vrr/opencode
2
0
Fork 0
mirror of https://github.com/anomalyco/opencode.git synced 2026-08-04 11:19:51 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6

feat(console): connect named go subscribers

Browse source
This commit is contained in:
vprdev 2026-08-03 11:56:34 -04:00
parent d7ee68a41d
commit 86e4631993
28 changed files with 370 additions and 4197 deletions
Download patch file Download diff file Expand all files Collapse all files

24
packages/console/app/src/i18n/en.ts
View file

@ -265,7 +265,8 @@ export const dict = {
"go.cta.text": "Subscribe to Go",
"go.cta.price": "$10/month",
"go.cta.promo": "$5 first month",
"go.pricing.body": "Use with any agent. $5 first month, then $10/month. Top up credit if needed. Cancel any time.",
"go.pricing.body":
"For a named OpenCode subscriber. $5 first month, then $10/month. Service accounts are not eligible. Cancel any time.",
"go.graph.free": "Free",
"go.graph.freePill": "Big Pickle and free models",
"go.graph.go": "Go",
@ -304,15 +305,15 @@ export const dict = {
"go.problem.item4":
"Includes Grok 4.5, GLM-5.2, GLM-5.1, Kimi K3, Kimi K2.7 Code, Kimi K2.6, MiMo-V2.5-Pro, MiMo-V2.5, Qwen3.7 Max, Qwen3.7 Plus, Qwen3.6 Plus, MiniMax M2.7, MiniMax M3, DeepSeek V4 Pro, DeepSeek V4 Flash, and Hy3",
"go.how.title": "How Go works",
"go.how.body": "Go starts at $5 for your first month, then $10/month. You can use it with OpenCode or any agent.",
"go.how.step1.title": "Create an account",
"go.how.step1.beforeLink": "follow the",
"go.how.step1.link": "setup instructions",
"go.how.step2.title": "Subscribe to Go",
"go.how.step2.link": "$5 first month",
"go.how.step2.afterLink": "then $10/month with generous limits",
"go.how.body": "Go is available to a named subscriber using OpenCode. No API key needs to be copied.",
"go.how.step1.title": "Subscribe to Go",
"go.how.step1.beforeLink": "in",
"go.how.step1.link": "OpenCode Console",
"go.how.step2.title": "Connect OpenCode",
"go.how.step2.link": "run opencode console login",
"go.how.step2.afterLink": "and authorize the device in your browser",
"go.how.step3.title": "Start coding",
"go.how.step3.body": "with reliable access to open-source models",
"go.how.step3.body": "with the opencode provider",
"go.privacy.title": "Your privacy is important to us",
"go.privacy.body":
"The plan is designed primarily for international users, with models hosted in the US, EU, and Singapore for stable global access.",
@ -346,8 +347,9 @@ export const dict = {
"go.faq.a6": "If you need more usage, you can top up credit in your account.",
"go.faq.q7": "Can I cancel?",
"go.faq.a7": "Yes, you can cancel any time.",
"go.faq.q8": "Can I use Go with other coding agents?",
"go.faq.a8": "Yes, you can use Go with any agent. Follow the setup instructions in your preferred coding agent.",
"go.faq.q8": "Who can use Go?",
"go.faq.a8":
"Go is available to the named subscriber through OpenCode. Other coding agents and service accounts are not eligible.",
"go.faq.q9": "What is the difference between free models and Go?",
"go.faq.a9":

53
packages/console/app/src/i18n/index.ts
View file

@ -23,23 +23,42 @@ export type Dict = Record<Key, string>
const base = en satisfies Dict
const goFallback = {
"go.pricing.body": base["go.pricing.body"],
"go.how.body": base["go.how.body"],
"go.how.step1.title": base["go.how.step1.title"],
"go.how.step1.beforeLink": base["go.how.step1.beforeLink"],
"go.how.step1.link": base["go.how.step1.link"],
"go.how.step2.title": base["go.how.step2.title"],
"go.how.step2.link": base["go.how.step2.link"],
"go.how.step2.afterLink": base["go.how.step2.afterLink"],
"go.how.step3.title": base["go.how.step3.title"],
"go.how.step3.body": base["go.how.step3.body"],
"go.faq.q8": base["go.faq.q8"],
"go.faq.a8": base["go.faq.a8"],
}
function withGoFallback(dict: Dict): Dict {
return { ...dict, ...goFallback }
}
export function i18n(locale: Locale): Dict {
if (locale === "en") return base
if (locale === "zh") return { ...base, ...zh }
if (locale === "zht") return { ...base, ...zht }
if (locale === "ko") return { ...base, ...ko }
if (locale === "de") return { ...base, ...de }
if (locale === "es") return { ...base, ...es }
if (locale === "fr") return { ...base, ...fr }
if (locale === "it") return { ...base, ...it }
if (locale === "da") return { ...base, ...da }
if (locale === "ja") return { ...base, ...ja }
if (locale === "pl") return { ...base, ...pl }
if (locale === "ru") return { ...base, ...ru }
if (locale === "uk") return { ...base, ...uk }
if (locale === "ar") return { ...base, ...ar }
if (locale === "no") return { ...base, ...no }
if (locale === "br") return { ...base, ...br }
if (locale === "th") return { ...base, ...th }
return { ...base, ...tr }
if (locale === "zh") return withGoFallback({ ...base, ...zh })
if (locale === "zht") return withGoFallback({ ...base, ...zht })
if (locale === "ko") return withGoFallback({ ...base, ...ko })
if (locale === "de") return withGoFallback({ ...base, ...de })
if (locale === "es") return withGoFallback({ ...base, ...es })
if (locale === "fr") return withGoFallback({ ...base, ...fr })
if (locale === "it") return withGoFallback({ ...base, ...it })
if (locale === "da") return withGoFallback({ ...base, ...da })
if (locale === "ja") return withGoFallback({ ...base, ...ja })
if (locale === "pl") return withGoFallback({ ...base, ...pl })
if (locale === "ru") return withGoFallback({ ...base, ...ru })
if (locale === "uk") return withGoFallback({ ...base, ...uk })
if (locale === "ar") return withGoFallback({ ...base, ...ar })
if (locale === "no") return withGoFallback({ ...base, ...no })
if (locale === "br") return withGoFallback({ ...base, ...br })
if (locale === "th") return withGoFallback({ ...base, ...th })
return withGoFallback({ ...base, ...tr })
}

5
packages/console/app/src/routes/go/index.tsx
View file

@ -372,7 +372,7 @@ export default function Home() {
<span>[1]</span>
<div>
<strong>{i18n.t("go.how.step1.title")}</strong> - {i18n.t("go.how.step1.beforeLink")}{" "}
<a href={language.route("/docs/go/#how-it-works")} title={i18n.t("go.how.step1.link")}>
<a href="/console/go" title={i18n.t("go.how.step1.link")}>
{i18n.t("go.how.step1.link")}
</a>
</div>
@ -381,8 +381,7 @@ export default function Home() {
<span>[2]</span>
<div>
<strong>{i18n.t("go.how.step2.title")}</strong> -{" "}
<a href={language.route("/docs/go/#pricing")}>{i18n.t("go.how.step2.link")}</a>{" "}
{i18n.t("go.how.step2.afterLink")}
<a href="/v2/docs/go">{i18n.t("go.how.step2.link")}</a> {i18n.t("go.how.step2.afterLink")}
</div>
</li>
<li>

42
packages/core/src/plugin/provider/opencode.ts
View file

@ -14,7 +14,7 @@ import { Money } from "@opencode-ai/schema/money"
import { ConfigProviderOptionsV1 } from "../../v1/config/provider-options"
import { ConfigV1 } from "../../v1/config/config"
const defaultServer = "https://console.opencode.ai"
const defaultServer = "https://opencode.ai/console"
const clientID = "opencode-cli"
const methodID = Integration.MethodID.make("device")
const RemoteResponse = Schema.Struct({ config: ConfigV1.Info })
@ -111,7 +111,10 @@ export const OpencodePlugin = define<HttpClient.HttpClient | Bus.Service | Scope
integration.name = "OpenCode"
})
draft.method.update(oauth(http))
draft.method.update({ integrationID: "opencode", method: { type: "key", label: "API key (service account)" } })
draft.method.update({
integrationID: "opencode",
method: { type: "key", label: "API key (managed inference service account; not Go)" },
})
})
yield* load()
@ -214,6 +217,12 @@ function fetchProviders(http: HttpClient.HttpClient, value: CredentialValue) {
.pipe(
Effect.flatMap((response) => {
if (response.status === 404) return Effect.succeed(undefined)
if (response.status === 403) {
return Effect.fail(new Error("OpenCode Console access is forbidden for the selected organization"))
}
if (response.status < 200 || response.status >= 300) {
return Effect.fail(new Error(`OpenCode Console provider config failed with HTTP ${response.status}`))
}
return HttpClientResponse.filterStatusOk(response).pipe(
Effect.flatMap(HttpClientResponse.schemaBodyJson(RemoteResponse)),
Effect.map((remote) => remote.config.provider),
@ -296,8 +305,22 @@ function credential(http: HttpClient.HttpClient, server: string, token: typeof T
],
{ concurrency: 2 },
)
const org = orgs.toSorted((a, b) => a.name.localeCompare(b.name) || a.id.localeCompare(b.id))[0]
return Credential.OAuth.make({
if (orgs.length === 0) {
return yield* Effect.fail(
new Error(
"Your OpenCode Console account does not belong to an organization. Create or join one at https://opencode.ai/console, then try again.",
),
)
}
if (orgs.length > 1) {
return yield* Effect.fail(
new Error(
"Your OpenCode Console account belongs to multiple organizations. Organization selection is not supported yet; use an account with one organization, then try again.",
),
)
}
const org = orgs[0]
const value = Credential.OAuth.make({
type: "oauth" as const,
methodID,
access: token.access_token,
@ -307,10 +330,17 @@ function credential(http: HttpClient.HttpClient, server: string, token: typeof T
server,
accountID: user.id,
email: user.email,
orgID: org?.id,
orgName: org?.name,
orgID: org.id,
orgName: org.name,
},
})
const providers = yield* fetchProviders(http, value)
if (!providers) {
return yield* Effect.fail(
new Error("OpenCode Console did not return provider config for the selected organization"),
)
}
return value
})
}

198
packages/core/test/plugin/provider-opencode.test.ts
View file

@ -1,6 +1,7 @@
import { describe, expect } from "bun:test"
import { Money } from "@opencode-ai/schema/money"
import { Effect } from "effect"
import { HttpClient, HttpClientResponse } from "effect/unstable/http"
import { Catalog } from "@opencode-ai/core/catalog"
import { Credential } from "@opencode-ai/core/credential"
import { Bus } from "@opencode-ai/core/bus"
@ -87,11 +88,50 @@ describe("OpencodePlugin", () => {
type: "oauth",
label: "OpenCode Console account",
},
{ type: "key", label: "API key (service account)" },
{ type: "key", label: "API key (managed inference service account; not Go)" },
])
}),
)
it.effect("uses the canonical OpenCode Console server by default", () =>
Effect.gen(function* () {
const requests: string[] = []
const http = HttpClient.make((request) => {
requests.push(request.url)
return Effect.succeed(
HttpClientResponse.fromWeb(
request,
Response.json({
device_code: "device",
user_code: "user",
verification_uri_complete: "/verify",
expires_in: 60,
interval: 60,
}),
),
)
})
const plugin = yield* Plugin.Service
const host = yield* PluginHost.make(plugin)
const bus = yield* Bus.Service
const integration = yield* Integration.Service
yield* OpencodePlugin.effect(host).pipe(
Effect.provideService(Bus.Service, bus),
Effect.provideService(Integration.Service, integration),
Effect.provideService(HttpClient.HttpClient, http),
)
const attempt = yield* integration.oauth.connect({
integrationID: Integration.ID.make("opencode"),
methodID: Integration.MethodID.make("device"),
inputs: {},
})
yield* integration.oauth.cancel({ integrationID: Integration.ID.make("opencode"), attemptID: attempt.attemptID })
expect(requests).toEqual(["https://opencode.ai/console/auth/device/code"])
expect(attempt.url).toBe("https://opencode.ai/console/verify")
}),
)
it.live("uses a canonical custom server throughout device authorization", () =>
Effect.acquireUseRelease(
Effect.sync(() => {
@ -115,6 +155,9 @@ describe("OpencodePlugin", () => {
}
if (url.pathname.endsWith("/api/user")) return Response.json({ id: "user", email: "user@example.com" })
if (url.pathname.endsWith("/api/orgs")) return Response.json([{ id: "org", name: "Org" }])
if (url.pathname.endsWith("/api/config")) {
return Response.json({ config: { enterprise: { url: url.origin }, provider: {} } })
}
return new Response("Not found", { status: 404 })
},
})
@ -140,14 +183,165 @@ describe("OpencodePlugin", () => {
expect(requests).toContain("POST /console/auth/device/token")
expect(requests).toContain("GET /console/api/user")
expect(requests).toContain("GET /console/api/orgs")
expect(requests).toContain("GET /console/api/config")
expect((yield* (yield* Credential.Service).list(Integration.ID.make("opencode")))[0]?.value).toMatchObject({
metadata: { server: `${server.url.origin}/console` },
metadata: { server: `${server.url.origin}/console`, orgID: "org", orgName: "Org" },
})
}),
({ server }) => Effect.promise(() => server.stop(true)),
),
)
it.live("rejects device login without an organization", () =>
Effect.acquireUseRelease(
Effect.sync(() =>
Bun.serve({
port: 0,
fetch: (request) => {
const url = new URL(request.url)
if (url.pathname === "/auth/device/code") {
return Response.json({
device_code: "device",
user_code: "user",
verification_uri_complete: `${url.origin}/verify`,
expires_in: 60,
interval: 0,
})
}
if (url.pathname === "/auth/device/token") {
return Response.json({ access_token: "access", refresh_token: "refresh", expires_in: 600 })
}
if (url.pathname === "/api/user") return Response.json({ id: "user", email: "user@example.com" })
if (url.pathname === "/api/orgs") return Response.json([])
return new Response("Not found", { status: 404 })
},
}),
),
(server) =>
Effect.gen(function* () {
yield* addPlugin()
const integrations = yield* Integration.Service
const integrationID = Integration.ID.make("opencode")
const attempt = yield* integrations.oauth.connect({
integrationID,
methodID: Integration.MethodID.make("device"),
inputs: { server: server.url.origin },
})
const status = yield* eventually(
integrations.oauth.status({ integrationID, attemptID: attempt.attemptID }),
(value) => value.status !== "pending",
)
expect(status).toMatchObject({ status: "failed" })
if (status.status === "failed") expect(status.message).toContain("does not belong to an organization")
expect(yield* (yield* Credential.Service).list(integrationID)).toEqual([])
}),
(server) => Effect.promise(() => server.stop(true)),
),
)
it.live("rejects device login with multiple organizations", () =>
Effect.acquireUseRelease(
Effect.sync(() =>
Bun.serve({
port: 0,
fetch: (request) => {
const url = new URL(request.url)
if (url.pathname === "/auth/device/code") {
return Response.json({
device_code: "device",
user_code: "user",
verification_uri_complete: `${url.origin}/verify`,
expires_in: 60,
interval: 0,
})
}
if (url.pathname === "/auth/device/token") {
return Response.json({ access_token: "access", refresh_token: "refresh", expires_in: 600 })
}
if (url.pathname === "/api/user") return Response.json({ id: "user", email: "user@example.com" })
if (url.pathname === "/api/orgs") {
return Response.json([
{ id: "org-b", name: "Beta" },
{ id: "org-a", name: "Alpha" },
])
}
return new Response("Not found", { status: 404 })
},
}),
),
(server) =>
Effect.gen(function* () {
yield* addPlugin()
const integrations = yield* Integration.Service
const integrationID = Integration.ID.make("opencode")
const attempt = yield* integrations.oauth.connect({
integrationID,
methodID: Integration.MethodID.make("device"),
inputs: { server: server.url.origin },
})
const status = yield* eventually(
integrations.oauth.status({ integrationID, attemptID: attempt.attemptID }),
(value) => value.status !== "pending",
)
expect(status).toMatchObject({ status: "failed" })
if (status.status === "failed") expect(status.message).toContain("multiple organizations")
expect(yield* (yield* Credential.Service).list(integrationID)).toEqual([])
}),
(server) => Effect.promise(() => server.stop(true)),
),
)
it.live("does not complete device login before provider config loads", () =>
Effect.acquireUseRelease(
Effect.sync(() =>
Bun.serve({
port: 0,
fetch: (request) => {
const url = new URL(request.url)
if (url.pathname === "/auth/device/code") {
return Response.json({
device_code: "device",
user_code: "user",
verification_uri_complete: `${url.origin}/verify`,
expires_in: 60,
interval: 0,
})
}
if (url.pathname === "/auth/device/token") {
return Response.json({ access_token: "access", refresh_token: "refresh", expires_in: 600 })
}
if (url.pathname === "/api/user") return Response.json({ id: "user", email: "user@example.com" })
if (url.pathname === "/api/orgs") return Response.json([{ id: "org", name: "Org" }])
if (url.pathname === "/api/config") return new Response("Forbidden", { status: 403 })
return new Response("Not found", { status: 404 })
},
}),
),
(server) =>
Effect.gen(function* () {
yield* addPlugin()
const integrations = yield* Integration.Service
const integrationID = Integration.ID.make("opencode")
const attempt = yield* integrations.oauth.connect({
integrationID,
methodID: Integration.MethodID.make("device"),
inputs: { server: server.url.origin },
})
const status = yield* eventually(
integrations.oauth.status({ integrationID, attemptID: attempt.attemptID }),
(value) => value.status !== "pending",
)
expect(status).toMatchObject({ status: "failed" })
if (status.status === "failed") expect(status.message).toContain("forbidden for the selected organization")
expect(yield* (yield* Credential.Service).list(integrationID)).toEqual([])
}),
(server) => Effect.promise(() => server.stop(true)),
),
)
it.effect("rejects non-HTTP OpenCode servers", () =>
Effect.gen(function* () {
yield* addPlugin()

224
packages/web/src/content/docs/ar/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: اشتراك منخفض التكلفة لنماذج البرمجة المفتوحة.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go هو اشتراك منخفض التكلفة — **$5 للشهر الأول**، ثم **$10/شهريًا** — يمنحك وصولًا موثوقًا إلى نماذج البرمجة المفتوحة الشائعة.
يعمل Go مثل أي مزود آخر في OpenCode. تشترك في OpenCode Go وتحصل على مفتاح API الخاص بك. وهو **اختياري تمامًا**، ولا تحتاج إلى استخدامه لاستخدام OpenCode.
صُمّم أساسًا للمستخدمين الدوليين، مع استضافة النماذج في الولايات المتحدة والاتحاد الأوروبي وسنغافورة لضمان وصول عالمي مستقر.
---
## الخلفية
أصبحت النماذج المفتوحة جيدة جدًا. فهي الآن تقترب في الأداء من النماذج المملوكة في مهام البرمجة. وبما أن كثيرًا من المزودين يمكنهم تقديمها بشكل تنافسي، فهي عادةً أقل تكلفة بكثير.
ومع ذلك، قد يكون الحصول على وصول موثوق إليها بزمن استجابة منخفض أمرًا صعبًا. إذ يختلف المزودون في الجودة والتوافر.
:::tip
اختبرنا مجموعة مختارة من النماذج والمزودين الذين يعملون جيدًا مع OpenCode.
:::
لمعالجة ذلك، قمنا بعدة أمور:
1. اختبرنا مجموعة مختارة من النماذج المفتوحة وتحدثنا مع فرقها حول أفضل طريقة لتشغيلها.
2. ثم عملنا مع عدد قليل من المزودين للتأكد من تقديم هذه النماذج بالشكل الصحيح.
3. وأخيرًا، أجرينا مقارنات معيارية لمزيج النموذج/المزود، وتوصلنا إلى قائمة نشعر بالثقة في التوصية بها.
يمنحك OpenCode Go الوصول إلى هذه النماذج مقابل **$5 للشهر الأول**، ثم **$10/شهريًا**.
---
## كيف يعمل
يعمل OpenCode Go مثل أي مزود آخر في OpenCode.
1. تسجّل الدخول إلى **<a href={console}>OpenCode Zen</a>**، وتشترك في Go، ثم تنسخ مفتاح API الخاص بك.
2. تشغّل الأمر `/connect` في TUI، وتختار `OpenCode Go`، ثم تلصق مفتاح API الخاص بك.
3. شغّل `/models` في TUI لرؤية قائمة النماذج المتاحة عبر Go.
:::note
يمكن لعضو واحد فقط في كل workspace الاشتراك في OpenCode Go.
:::
تشمل قائمة النماذج الحالية:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
قد تتغير قائمة النماذج مع استمرارنا في اختبار نماذج جديدة وإضافتها.
---
## حدود الاستخدام
يتضمن OpenCode Go الحدود التالية:
- **حد 5 ساعات** — استخدام بقيمة $12
- **الحد الأسبوعي** — استخدام بقيمة $30
- **الحد الشهري** — استخدام بقيمة $60
تُحدَّد الحدود بالقيمة بالدولار. وهذا يعني أن عدد طلباتك الفعلي يعتمد على النموذج الذي تستخدمه. تتيح النماذج الأقل تكلفة مثل DeepSeek V4 Flash عددًا أكبر من الطلبات، بينما تتيح النماذج الأعلى تكلفة مثل GLM-5.2 عددًا أقل.
يوضح الجدول أدناه عددًا تقديريًا للطلبات بناءً على أنماط استخدام Go المعتادة:
| Model | الطلبات لكل 5 ساعات | الطلبات في الأسبوع | الطلبات في الشهر |
| ----------------- | ------------------- | ------------------ | ---------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
تستند التقديرات إلى أنماط الطلبات المرصودة:
- Grok 4.5 — ‏1,100 input، و71,500 cached، و220 output tokens لكل طلب
- GLM-5.2/5.1 — ‏700 input، و52,000 cached، و150 output tokens لكل طلب
- Kimi K3 — ‏1,050 input، و76,500 cached، و300 output tokens لكل طلب
- Kimi K2.7/K2.6 — ‏870 input، و55,000 cached، و200 output tokens لكل طلب
- DeepSeek V4 Pro — ‏750 input، و82,000 cached، و290 output tokens لكل طلب
- DeepSeek V4 Flash — ‏790 input، و68,000 cached، و280 output tokens لكل طلب
- MiniMax M3 — ‏510 input، و56,000 cached، و190 output tokens لكل طلب
- MiniMax M2.7 — ‏300 input، و55,000 cached، و125 output tokens لكل طلب
- Qwen3.7 Max — ‏420 input، و66,000 cached، و200 output tokens لكل طلب
- Qwen3.7 Plus — ‏500 input، و57,000 cached، و190 output tokens لكل طلب
- Qwen3.6 Plus — ‏500 input، و57,000 cached، و190 output tokens لكل طلب
- Hy3 — ‏830 input، و71,500 cached، و295 output tokens لكل طلب
- MiMo-V2.5 — ‏830 input، و71,500 cached، و295 output tokens لكل طلب
- MiMo-V2.5-Pro — ‏790 input، و86,000 cached، و305 output tokens لكل طلب
تستند التقديرات أيضًا إلى الأسعار التالية لكل 1M tokens والاستخدام الشهري المتضمن مع كل نموذج:
| النموذج | الإدخال | الإخراج | القراءة المخزنة | الكتابة المخزنة | الاستخدام |
| ---------------------------- | ------- | ------- | --------------- | --------------- | --------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
يمكنك تتبّع استخدامك الحالي في **<a href={console}>console</a>**.
:::tip
إذا وصلت إلى حد الاستخدام، يمكنك الاستمرار في استخدام النماذج المجانية.
:::
قد تتغير حدود الاستخدام مع تعلّمنا من الاستخدام المبكر والملاحظات.
---
### الاستخدام بعد تجاوز الحدود
إذا كان لديك أيضًا رصيد في Zen balance، فيمكنك تفعيل خيار **Use balance** في console. عند تفعيله، سيعود Go إلى استخدام Zen balance الخاص بك بعد بلوغ حدود الاستخدام بدلًا من حظر الطلبات.
---
### لماذا يكون الاستخدام أقل لبعض النماذج
مع Go، تدفع $10 شهريًا، ونهدف إلى منحك استخدامًا بقيمة تعادل 6 أضعاف هذا المبلغ.
نحقق ذلك لمعظم النماذج من خلال الخصومات على الكميات الكبيرة وسعة GPU المحجوزة. ثم ننقل هذه الوفورات إليك من خلال معامل مضاعفة قدره 6.
بالنسبة إلى بعض النماذج، لم تتح لنا فرصة التفاوض على خصم أو استضافتها بتكلفة أقل، إما لأن النموذج جديد أو لأن أسعاره العامة مخفضة بالفعل.
بالنسبة إلى هذه النماذج، لا تزال تحصل على استخدام يزيد قليلًا عما ستحصل عليه إذا دفعت لمزودي النماذج مباشرةً؛ ولهذا يكون معامل مضاعفة استخدامها أقل في الجدول أعلاه.
---
## نقاط النهاية
يمكنك أيضًا الوصول إلى نماذج Go عبر نقاط نهاية API التالية.
| Model | Model ID | Endpoint | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
يستخدم [model id](/docs/config/#models) في إعدادات OpenCode لديك التنسيق `opencode-go/<model-id>`. على سبيل المثال، بالنسبة إلى Kimi K3، ستستخدم `opencode-go/kimi-k3` في إعداداتك.
---
### النماذج
يمكنك جلب القائمة الكاملة بالنماذج المتاحة وبياناتها الوصفية من:
```
https://opencode.ai/zen/go/v1/models
```
---
## الخصوصية
صُمّمت هذه الخطة أساسًا للمستخدمين الدوليين، مع استضافة النماذج في الولايات المتحدة والاتحاد الأوروبي وسنغافورة لضمان وصول عالمي مستقر. ويتّبع مزودونا سياسة عدم الاحتفاظ بالبيانات، ولا يستخدمون بياناتك لتدريب النماذج.
---
## الأهداف
أنشأنا OpenCode Go من أجل:
1. جعل البرمجة بالذكاء الاصطناعي **متاحة** لعدد أكبر من الأشخاص عبر اشتراك منخفض التكلفة.
2. توفير وصول **موثوق** إلى أفضل نماذج البرمجة المفتوحة.
3. تنسيق نماذج **مختبرة وخضعت لمقارنات معيارية** لاستخدامها مع وكلاء البرمجة.
4. **عدم فرض قيود ارتباط** من خلال السماح لك أيضًا باستخدام أي مزود آخر مع OpenCode.

238
packages/web/src/content/docs/bs/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Povoljna pretplata za otvorene modele za programiranje.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go je povoljna pretplata — **$5 za vaš prvi mjesec**, a zatim **$10/mjesečno** — koja vam pruža pouzdan pristup popularnim otvorenim modelima za programiranje.
Go radi kao bilo koji drugi provajder u OpenCode-u. Pretplatite se na OpenCode Go i
dobijete svoj API ključ. On je **potpuno opcionalan** i ne morate ga koristiti da
biste koristili OpenCode.
Dizajniran je prvenstveno za međunarodne korisnike, sa modelima hostovanim u SAD-u, EU i Singapuru za stabilan globalni pristup.
---
## Pozadina
Otvoreni modeli su postali zaista dobri. Sada dostižu performanse bliske
vlasničkim modelima za zadatke programiranja. A pošto ih mnogi provajderi mogu nuditi
konkurentno, obično su znatno jeftiniji.
Međutim, dobiti pouzdan pristup s niskom latencijom do njih može biti teško. Provajderi
variraju u pogledu kvaliteta i dostupnosti.
:::tip
Testirali smo odabranu grupu modela i provajdera koji dobro rade sa OpenCode-om.
:::
Da bismo to popravili, uradili smo nekoliko stvari:
1. Testirali smo odabranu grupu otvorenih modela i razgovarali s njihovim timovima o tome kako da ih
najbolje pokrenemo.
2. Zatim smo sarađivali s nekoliko provajdera kako bismo bili sigurni da se oni ispravno
poslužuju.
3. Na kraju smo benchmarkovali kombinaciju modela/provajdera i osmislili
listu koju rado preporučujemo.
OpenCode Go vam daje pristup ovim modelima za **$5 za vaš prvi mjesec**, a zatim **$10/mjesečno**.
---
## Kako funkcioniše
OpenCode Go radi kao bilo koji drugi provajder u OpenCode-u.
1. Prijavite se na **<a href={console}>OpenCode Zen</a>**, pretplatite se na Go i
kopirajte svoj API ključ.
2. Pokrenite komandu `/connect` u TUI-ju, odaberite `OpenCode Go` i zalijepite
svoj API ključ.
3. Pokrenite `/models` u TUI-ju da vidite listu modela dostupnih kroz Go.
:::note
Samo jedan član po radnom prostoru (workspace) može se pretplatiti na OpenCode Go.
:::
Trenutna lista modela uključuje:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Lista modela se može mijenjati dok testiramo i dodajemo nove.
---
## Ograničenja upotrebe
OpenCode Go uključuje sljedeća ograničenja:
- **Ograničenje od 5 sati** — $12 potrošnje
- **Sedmično ograničenje** — $30 potrošnje
- **Mjesečno ograničenje** — $60 potrošnje
Ograničenja su definisana u dolarskoj vrijednosti. To znači da vaš stvarni broj zahtjeva zavisi od modela koji koristite. Jeftiniji modeli poput DeepSeek V4 Flash omogućavaju više zahtjeva, dok skuplji modeli poput GLM-5.2 omogućavaju manje.
Tabela ispod pruža procijenjeni broj zahtjeva na osnovu tipičnih obrazaca korištenja Go pretplate:
| Model | zahtjeva na 5 sati | zahtjeva sedmično | zahtjeva mjesečno |
| ----------------- | ------------------ | ----------------- | ----------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Procjene se zasnivaju na zapaženim obrascima zahtjeva:
- Grok 4.5 — 1,100 ulaznih, 71,500 keširanih, 220 izlaznih tokena po zahtjevu
- GLM-5.2/5.1 — 700 ulaznih (input), 52,000 keširanih, 150 izlaznih (output) tokena po zahtjevu
- Kimi K3 — 1,050 ulaznih, 76,500 keširanih, 300 izlaznih tokena po zahtjevu
- Kimi K2.7/K2.6 — 870 ulaznih, 55,000 keširanih, 200 izlaznih tokena po zahtjevu
- DeepSeek V4 Pro — 750 ulaznih, 82,000 keširanih, 290 izlaznih tokena po zahtjevu
- DeepSeek V4 Flash — 790 ulaznih, 68,000 keširanih, 280 izlaznih tokena po zahtjevu
- MiniMax M3 — 510 ulaznih, 56,000 keširanih, 190 izlaznih tokena po zahtjevu
- MiniMax M2.7 — 300 ulaznih, 55,000 keširanih, 125 izlaznih tokena po zahtjevu
- Qwen3.7 Max — 420 ulaznih, 66,000 keširanih, 200 izlaznih tokena po zahtjevu
- Qwen3.7 Plus — 500 ulaznih, 57,000 keširanih, 190 izlaznih tokena po zahtjevu
- Qwen3.6 Plus — 500 ulaznih, 57,000 keširanih, 190 izlaznih tokena po zahtjevu
- Hy3 — 830 ulaznih, 71,500 keširanih, 295 izlaznih tokena po zahtjevu
- MiMo-V2.5 — 830 ulaznih, 71,500 keširanih, 295 izlaznih tokena po zahtjevu
- MiMo-V2.5-Pro — 790 ulaznih, 86,000 keširanih, 305 izlaznih tokena po zahtjevu
Procjene se također zasnivaju na sljedećim cijenama po 1M tokena i mjesečnoj potrošnji uključenoj uz svaki model:
| Model | Input | Output | Cached Read | Cached Write | Potrošnja |
| ---------------------------- | ------ | ------ | ----------- | ------------ | --------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Svoju trenutnu potrošnju možete pratiti u **<a href={console}>konzoli</a>**.
:::tip
Ako dostignete ograničenje upotrebe, možete nastaviti koristiti besplatne modele.
:::
Ograničenja upotrebe mogu se promijeniti kako budemo učili iz rane upotrebe i povratnih informacija.
---
### Upotreba preko ograničenja
Ako također imate kredite na svom Zen balansu, možete omogućiti **Use balance**
opciju u konzoli. Kada je omogućeno, Go će preći na vaš Zen balans
nakon što dostignete ograničenja upotrebe umjesto blokiranja zahtjeva.
---
### Zašto neki modeli imaju manju uključenu potrošnju
Uz Go plaćate $10 mjesečno, a cilj nam je omogućiti vam potrošnju šest puta veću od tog iznosa.
Za većinu modela to postižemo količinskim popustima i rezervisanim GPU kapacitetom. Tu uštedu zatim prenosimo na vas primjenom faktora šest.
Za neke modele još nismo imali priliku dogovoriti popust ili ih hostovati po nižoj cijeni, bilo zato što su modeli novi ili zato što su njihove javno objavljene cijene već snižene.
Za ove modele i dalje dobijate malo više nego da direktno plaćate provajderima modela; zato je njihov faktor potrošnje niži u gornjoj tabeli.
---
## Endpointi
Također možete pristupiti Go modelima putem sljedećih API endpointa.
| Model | Model ID | Endpoint | AI SDK Paket |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
[Model id](/docs/config/#models) u vašoj OpenCode konfiguraciji
koristi format `opencode-go/<model-id>`. Na primjer, za Kimi K3, koristili biste
`opencode-go/kimi-k3` u svojoj konfiguraciji.
---
### Modeli
Pun spisak dostupnih modela i njihovih metapodataka možete preuzeti na:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privatnost
Plan je prvenstveno namijenjen međunarodnim korisnicima, a modeli su smješteni u US, EU i Singaporeu radi stabilnog globalnog pristupa. Naši pružaoci usluga primjenjuju politiku nultog zadržavanja podataka i ne koriste vaše podatke za treniranje modela.
---
## Ciljevi
Napravili smo OpenCode Go da bismo:
1. Učinili AI programiranje **dostupnim** većem broju ljudi putem povoljne pretplate.
2. Pružili **pouzdan** pristup najboljim otvorenim modelima za programiranje.
3. Odabrali modele koji su **testirani i benchmarkovani** za upotrebu od strane agenata za programiranje.
4. Osigurali da **nema zaključavanja (no lock-in)**, omogućavajući vam da uz OpenCode koristite i bilo kojeg drugog provajdera.

238
packages/web/src/content/docs/da/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Lavprisabonnement til åbne kodningsmodeller.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go er et lavprisabonnement — **$5 for din første måned**, derefter **$10/måned** — der giver dig pålidelig adgang til populære åbne kodningsmodeller.
Go fungerer som enhver anden udbyder i OpenCode. Du abonnerer på OpenCode Go og
får din API-nøgle. Det er **helt valgfrit**, og du behøver ikke at bruge det for at
bruge OpenCode.
Det er primært designet til internationale brugere, med modeller hostet i USA, EU og Singapore for at sikre stabil global adgang.
---
## Baggrund
Åbne modeller er blevet rigtig gode. De opnår nu en ydeevne, der er tæt på
proprietære modeller til kodningsopgaver. Og fordi mange udbydere kan levere dem
konkurrencedygtigt, er de som regel langt billigere.
Det kan dog være svært at få pålidelig adgang til dem med lav latenstid. Udbydere
varierer i kvalitet og tilgængelighed.
:::tip
Vi testede en udvalgt gruppe af modeller og udbydere, der fungerer godt med OpenCode.
:::
For at løse dette, gjorde vi et par ting:
1. Vi testede en udvalgt gruppe af åbne modeller og talte med deres teams om, hvordan man
bedst kører dem.
2. Derefter arbejdede vi med nogle få udbydere for at sikre, at disse blev leveret
korrekt.
3. Til sidst benchmarkede vi kombinationen af model og udbyder, og kom frem til
en liste, som vi trygt kan anbefale.
OpenCode Go giver dig adgang til disse modeller for **$5 for din første måned**, derefter **$10/måned**.
---
## Sådan fungerer det
OpenCode Go fungerer som enhver anden udbyder i OpenCode.
1. Du logger ind på **<a href={console}>OpenCode Zen</a>**, abonnerer på Go, og
kopierer din API-nøgle.
2. Du kører kommandoen `/connect` i TUI'en, vælger `OpenCode Go`, og indsætter
din API-nøgle.
3. Kør `/models` i TUI'en for at se listen over tilgængelige modeller gennem Go.
:::note
Kun ét medlem per arbejdsområde kan abonnere på OpenCode Go.
:::
Den nuværende liste over modeller inkluderer:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Listen over modeller kan ændre sig, efterhånden som vi tester og tilføjer nye.
---
## Forbrugsgrænser
OpenCode Go inkluderer følgende grænser:
- **5-timers grænse** — forbrug for $12
- **Ugentlig grænse** — forbrug for $30
- **Månedlig grænse** — forbrug for $60
Grænserne er defineret i dollarværdi. Det betyder, at dit faktiske antal anmodninger afhænger af den model, du bruger. Billigere modeller som DeepSeek V4 Flash tillader flere anmodninger, mens dyrere modeller som GLM-5.2 tillader færre.
Tabellen nedenfor giver et estimeret antal anmodninger baseret på typiske Go-forbrugsmønstre:
| Model | anmodninger pr. 5 timer | anmodninger pr. uge | anmodninger pr. måned |
| ----------------- | ----------------------- | ------------------- | --------------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Estimaterne er baseret på observerede anmodningsmønstre:
- Grok 4.5 — 1.100 input, 71.500 cachelagrede, 220 output-tokens pr. anmodning
- GLM-5.2/5.1 — 700 input, 52.000 cachelagrede, 150 output-tokens pr. anmodning
- Kimi K3 — 1.050 input, 76.500 cachelagrede, 300 output-tokens pr. anmodning
- Kimi K2.7/K2.6 — 870 input, 55.000 cachelagrede, 200 output-tokens pr. anmodning
- DeepSeek V4 Pro — 750 input, 82.000 cachelagrede, 290 output-tokens pr. anmodning
- DeepSeek V4 Flash — 790 input, 68.000 cachelagrede, 280 output-tokens pr. anmodning
- MiniMax M3 — 510 input, 56.000 cachelagrede, 190 output-tokens pr. anmodning
- MiniMax M2.7 — 300 input, 55.000 cachelagrede, 125 output-tokens pr. anmodning
- Qwen3.7 Max — 420 input, 66.000 cachelagrede, 200 output-tokens pr. anmodning
- Qwen3.7 Plus — 500 input, 57.000 cachelagrede, 190 output-tokens pr. anmodning
- Qwen3.6 Plus — 500 input, 57.000 cachelagrede, 190 output-tokens pr. anmodning
- Hy3 — 830 input, 71.500 cachelagrede, 295 output-tokens pr. anmodning
- MiMo-V2.5 — 830 input, 71.500 cachelagrede, 295 output-tokens pr. anmodning
- MiMo-V2.5-Pro — 790 input, 86.000 cachelagrede, 305 output-tokens pr. anmodning
Estimaterne er også baseret på følgende priser pr. 1M tokens og det månedlige forbrug, der er inkluderet med hver model:
| Model | Input | Output | Cached Read | Cached Write | Forbrug |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Du kan spore dit nuværende forbrug i **<a href={console}>konsollen</a>**.
:::tip
Hvis du når forbrugsgrænsen, kan du fortsætte med at bruge de gratis modeller.
:::
Forbrugsgrænser kan ændre sig, efterhånden som vi lærer af tidlig brug og feedback.
---
### Forbrug ud over grænserne
Hvis du også har kredit på din Zen-saldo, kan du aktivere indstillingen **Use balance**
i konsollen. Når den er aktiveret, vil Go falde tilbage på din Zen-saldo,
når du har nået dine forbrugsgrænser, i stedet for at blokere anmodninger.
---
### Hvorfor nogle modeller har lavere forbrug
Med Go betaler du $10/måned, og vi sigter mod at give dig 6x så meget i forbrug.
For de fleste modeller gør vi dette muligt gennem mængderabatter og reserveret GPU-kapacitet. Vi giver dig derefter disse besparelser videre gennem 6x-multiplikatoren.
For nogle modeller har vi endnu ikke haft mulighed for at forhandle en rabat eller hoste dem til en lavere pris, enten fordi modellen er ny, eller fordi dens offentlige pris allerede er nedsat.
Med disse modeller får du stadig lidt mere, end hvis du betalte modeludbyderne direkte. Derfor er deres forbrugsmultiplikator lavere i tabellen ovenfor.
---
## Endpoints
Du kan også få adgang til Go-modeller gennem følgende API-endpoints.
| Model | Model ID | Endpoint | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
Dit [model id](/docs/config/#models) i din OpenCode config
bruger formatet `opencode-go/<model-id>`. For eksempel for Kimi K3, vil du
bruge `opencode-go/kimi-k3` i din config.
---
### Modeller
Du kan hente den fulde liste over tilgængelige modeller og deres metadata fra:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privatliv
Planen er primært designet til internationale brugere med modeller hostet i US, EU og Singapore for stabil global adgang. Vores udbydere følger en zero-retention-policy og bruger ikke dine data til modeltræning.
---
## Mål
Vi skabte OpenCode Go for at:
1. Gøre AI-kodning **tilgængelig** for flere mennesker med et lavprisabonnement.
2. Give **pålidelig** adgang til de bedste åbne kodningsmodeller.
3. Kuratere modeller, der er **testet og benchmarkede** til brug med kodningsagenter.
4. Undgå **lock-in** ved også at lade dig bruge enhver anden udbyder med OpenCode.

226
packages/web/src/content/docs/de/go.mdx
View file

@ -1,226 +0,0 @@
---
title: Go
description: Kostengünstiges Abonnement für offene Coding-Modelle.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go ist ein kostengünstiges Abonnement — **5 $ für deinen ersten Monat**, danach **10 $/Monat** —, das dir zuverlässigen Zugriff auf beliebte offene Coding-Modelle bietet.
Go funktioniert wie jeder andere Provider in OpenCode. Du abonnierst OpenCode Go und
erhältst deinen API-Key. Es ist **völlig optional** und du musst es nicht nutzen, um
OpenCode zu verwenden.
Es wurde primär für internationale Nutzer entwickelt, wobei die Modelle für einen stabilen weltweiten Zugriff in den USA, der EU und Singapur gehostet werden.
---
## Hintergrund
Offene Modelle sind richtig gut geworden. Sie erreichen bei Coding-Aufgaben mittlerweile eine Leistung, die nahe an proprietäre Modelle herankommt. Und da viele Provider sie wettbewerbsfähig anbieten können, sind sie in der Regel viel günstiger.
Allerdings kann es schwierig sein, zuverlässigen Zugriff mit geringer Latenz auf sie zu erhalten. Provider variieren in Qualität und Verfügbarkeit.
:::tip
Wir haben eine ausgewählte Gruppe von Modellen und Providern getestet, die gut mit OpenCode funktionieren.
:::
Um dies zu beheben, haben wir einige Dinge getan:
1. Wir haben eine ausgewählte Gruppe offener Modelle getestet und mit ihren Teams darüber gesprochen, wie man sie am besten ausführt.
2. Anschließend haben wir mit einigen Providern zusammengearbeitet, um sicherzustellen, dass diese korrekt bereitgestellt werden.
3. Zuletzt haben wir die Kombination aus Modell und Provider einem Benchmark unterzogen und eine Liste erstellt, die wir mit gutem Gewissen empfehlen können.
OpenCode Go bietet dir Zugriff auf diese Modelle für **5 $ im ersten Monat**, danach **10 $/Monat**.
---
## Wie es funktioniert
OpenCode Go funktioniert wie jeder andere Provider in OpenCode.
1. Du meldest dich bei **<a href={console}>OpenCode Zen</a>** an, abonnierst Go und kopierst deinen API-Key.
2. Du führst den Befehl `/connect` in der TUI aus, wählst `OpenCode Go` und fügst deinen API-Key ein.
3. Führe `/models` in der TUI aus, um die Liste der über Go verfügbaren Modelle zu sehen.
:::note
Nur ein Mitglied pro Workspace kann OpenCode Go abonnieren.
:::
Die aktuelle Liste der Modelle umfasst:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Die Liste der Modelle kann sich ändern, während wir neue testen und hinzufügen.
---
## Nutzungslimits
OpenCode Go beinhaltet die folgenden Limits:
- **5-Stunden-Limit** — 12 $ Nutzung
- **Wöchentliches Limit** — 30 $ Nutzung
- **Monatliches Limit** — 60 $ Nutzung
Limits sind in Dollarwerten definiert. Das bedeutet, dass die tatsächliche Anzahl deiner Anfragen von dem von dir genutzten Modell abhängt. Günstigere Modelle wie DeepSeek V4 Flash erlauben mehr Anfragen, während teurere Modelle wie GLM-5.2 weniger erlauben.
Die folgende Tabelle zeigt eine geschätzte Anzahl von Anfragen basierend auf typischen Go-Nutzungsmustern:
| Model | Anfragen pro 5 Stunden | Anfragen pro Woche | Anfragen pro Monat |
| ----------------- | ---------------------- | ------------------ | ------------------ |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Die Schätzungen basieren auf beobachteten Anfragemustern:
- Grok 4.5 — 1.100 Input-, 71.500 Cached-, 220 Output-Tokens pro Anfrage
- GLM-5.2/5.1 — 700 Input-, 52.000 Cached-, 150 Output-Tokens pro Anfrage
- Kimi K3 — 1.050 Input-, 76.500 Cached-, 300 Output-Tokens pro Anfrage
- Kimi K2.7/K2.6 — 870 Input-, 55.000 Cached-, 200 Output-Tokens pro Anfrage
- DeepSeek V4 Pro — 750 Input-, 82.000 Cached-, 290 Output-Tokens pro Anfrage
- DeepSeek V4 Flash — 790 Input-, 68.000 Cached-, 280 Output-Tokens pro Anfrage
- MiniMax M3 — 510 Input-, 56.000 Cached-, 190 Output-Tokens pro Anfrage
- MiniMax M2.7 — 300 Input-, 55.000 Cached-, 125 Output-Tokens pro Anfrage
- Qwen3.7 Max — 420 Input-, 66.000 Cached-, 200 Output-Tokens pro Anfrage
- Qwen3.7 Plus — 500 Input-, 57.000 Cached-, 190 Output-Tokens pro Anfrage
- Qwen3.6 Plus — 500 Input-, 57.000 Cached-, 190 Output-Tokens pro Anfrage
- Hy3 — 830 Input-, 71.500 Cached-, 295 Output-Tokens pro Anfrage
- MiMo-V2.5 — 830 Input-, 71.500 Cached-, 295 Output-Tokens pro Anfrage
- MiMo-V2.5-Pro — 790 Input-, 86.000 Cached-, 305 Output-Tokens pro Anfrage
Die Schätzungen basieren außerdem auf den folgenden Preisen pro 1M Tokens und der monatlichen Nutzung, die bei jedem Modell enthalten ist:
| Model | Input | Output | Cached Read | Cached Write | Nutzung |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Du kannst deine aktuelle Nutzung in der **<a href={console}>Console</a>** verfolgen.
:::tip
Wenn du das Nutzungslimit erreichst, kannst du weiterhin die kostenlosen Modelle verwenden.
:::
Die Nutzungslimits können sich ändern, während wir aus der frühen Nutzung und dem Feedback lernen.
---
### Nutzung über die Limits hinaus
Wenn du auch Guthaben auf deinem Zen-Konto hast, kannst du in der Console die Option **Use balance** aktivieren. Wenn diese aktiviert ist, greift Go auf dein Zen-Guthaben zurück, nachdem du deine Nutzungslimits erreicht hast, anstatt Anfragen zu blockieren.
---
### Warum einige Modelle weniger Nutzung bieten
Mit Go zahlst du $10/Monat, und unser Ziel ist, dir dafür das Sechsfache dieses Betrags als Nutzungsguthaben zu bieten.
Bei den meisten Modellen ermöglichen wir dies durch Mengenrabatte und reservierte GPU-Kapazität. Diese Ersparnisse geben wir dann über den 6x-Multiplikator an dich weiter.
Bei einigen Modellen hatten wir bisher keine Gelegenheit, einen Rabatt auszuhandeln oder sie kostengünstiger zu hosten, entweder weil das jeweilige Modell neu oder sein öffentlicher Preis bereits rabattiert ist.
Bei diesen Modellen erhältst du immer noch etwas mehr, als wenn du die Modellanbieter direkt bezahlen würdest. Deshalb ist ihr Nutzungsmultiplikator in der Tabelle oben niedriger.
---
## Endpunkte
Du kannst auf die Go-Modelle auch über die folgenden API-Endpunkte zugreifen.
| Modell | Modell-ID | Endpunkt | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
Die [Modell-ID](/docs/config/#models) in deiner OpenCode Config verwendet das Format `opencode-go/<model-id>`. Für Kimi K3 würdest du beispielsweise `opencode-go/kimi-k3` in deiner Config verwenden.
---
### Models
Du kannst die vollständige Liste der verfügbaren Modelle und ihrer Metadaten hier abrufen:
```
https://opencode.ai/zen/go/v1/models
```
---
## Datenschutz
Der Plan ist in erster Linie für internationale Nutzer konzipiert, mit in US, EU und Singapore gehosteten Modellen für einen stabilen weltweiten Zugriff. Unsere Anbieter befolgen eine Zero-Retention-Richtlinie und verwenden Ihre Daten nicht für das Modelltraining.
---
## Ziele
Wir haben OpenCode Go entwickelt, um:
1. AI Coding durch ein kostengünstiges Abonnement für mehr Menschen **zugänglich** zu machen.
2. **Zuverlässigen** Zugriff auf die besten offenen Coding-Modelle zu bieten.
3. Modelle zu kuratieren, die für den Einsatz als Coding Agent **getestet und einem Benchmark unterzogen** wurden.
4. **Keinen Lock-in-Effekt** zu haben, indem wir dir ermöglichen, auch jeden anderen Provider mit OpenCode zu nutzen.

238
packages/web/src/content/docs/es/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Suscripción de bajo costo para modelos abiertos de programación.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go es una suscripción de bajo costo — **$5 por tu primer mes**, luego **$10/mes** — que te brinda acceso confiable a modelos abiertos de programación populares.
Go funciona como cualquier otro proveedor en OpenCode. Te suscribes a OpenCode Go y
obtienes tu API key. Es **completamente opcional** y no necesitas usarlo para
usar OpenCode.
Está diseñado principalmente para usuarios internacionales, con modelos alojados en EE. UU., la UE y Singapur para un acceso global estable.
---
## Antecedentes
Los modelos abiertos se han vuelto realmente buenos. Ahora alcanzan un rendimiento cercano a
los modelos propietarios en tareas de programación. Y como muchos proveedores pueden servirlos
de manera competitiva, suelen ser mucho más económicos.
Sin embargo, obtener un acceso confiable y de baja latencia a ellos puede ser difícil. Los proveedores
varían en calidad y disponibilidad.
:::tip
Probamos un grupo selecto de modelos y proveedores que funcionan bien con OpenCode.
:::
Para solucionar esto, hicimos un par de cosas:
1. Probamos un grupo selecto de modelos abiertos y hablamos con sus equipos sobre cómo
ejecutarlos mejor.
2. Luego trabajamos con algunos proveedores para asegurarnos de que se sirvieran
correctamente.
3. Finalmente, evaluamos el rendimiento de la combinación del modelo/proveedor y elaboramos
una lista que nos sentimos seguros de recomendar.
OpenCode Go te da acceso a estos modelos por **$5 por tu primer mes**, luego **$10/mes**.
---
## Cómo funciona
OpenCode Go funciona como cualquier otro proveedor en OpenCode.
1. Inicias sesión en **<a href={console}>OpenCode Zen</a>**, te suscribes a Go y
copias tu API key.
2. Ejecutas el comando `/connect` en la TUI, seleccionas `OpenCode Go` y pegas
tu API key.
3. Ejecutas `/models` en la TUI para ver la lista de modelos disponibles a través de Go.
:::note
Solo un miembro por espacio de trabajo puede suscribirse a OpenCode Go.
:::
La lista actual de modelos incluye:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
La lista de modelos puede cambiar a medida que probamos y agregamos otros nuevos.
---
## Límites de uso
OpenCode Go incluye los siguientes límites:
- **Límite de 5 horas** — $12 de uso
- **Límite semanal** — $30 de uso
- **Límite mensual** — $60 de uso
Los límites se definen en valor en dólares. Esto significa que tu cantidad real de peticiones depende del modelo que uses. Los modelos más económicos como DeepSeek V4 Flash permiten más peticiones, mientras que los modelos de mayor costo como GLM-5.2 permiten menos.
La siguiente tabla proporciona una cantidad estimada de peticiones basada en los patrones típicos de uso de Go:
| Model | peticiones por 5 horas | peticiones por semana | peticiones por mes |
| ----------------- | ---------------------- | --------------------- | ------------------ |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Las estimaciones se basan en los patrones de peticiones observados:
- Grok 4.5 — 1,100 tokens de entrada, 71,500 en caché, 220 tokens de salida por petición
- GLM-5.2/5.1 — 700 tokens de entrada, 52,000 en caché, 150 tokens de salida por petición
- Kimi K3 — 1,050 tokens de entrada, 76,500 en caché, 300 tokens de salida por petición
- Kimi K2.7/K2.6 — 870 tokens de entrada, 55,000 en caché, 200 tokens de salida por petición
- DeepSeek V4 Pro — 750 tokens de entrada, 82,000 en caché, 290 tokens de salida por petición
- DeepSeek V4 Flash — 790 tokens de entrada, 68,000 en caché, 280 tokens de salida por petición
- MiniMax M3 — 510 tokens de entrada, 56,000 en caché, 190 tokens de salida por petición
- MiniMax M2.7 — 300 tokens de entrada, 55,000 en caché, 125 tokens de salida por petición
- Qwen3.7 Max — 420 tokens de entrada, 66,000 en caché, 200 tokens de salida por petición
- Qwen3.7 Plus — 500 tokens de entrada, 57,000 en caché, 190 tokens de salida por petición
- Qwen3.6 Plus — 500 tokens de entrada, 57,000 en caché, 190 tokens de salida por petición
- Hy3 — 830 tokens de entrada, 71,500 en caché, 295 tokens de salida por petición
- MiMo-V2.5 — 830 tokens de entrada, 71,500 en caché, 295 tokens de salida por petición
- MiMo-V2.5-Pro — 790 tokens de entrada, 86,000 en caché, 305 tokens de salida por petición
Las estimaciones también se basan en los siguientes precios por 1M tokens y en el uso mensual incluido con cada modelo:
| Modelo | Entrada | Salida | Lectura en caché | Escritura en caché | Uso |
| ---------------------------- | ------- | ------ | ---------------- | ------------------ | --- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Puedes realizar un seguimiento de tu uso actual en la **<a href={console}>consola</a>**.
:::tip
Si alcanzas el límite de uso, puedes continuar usando los modelos gratuitos.
:::
Los límites de uso pueden cambiar a medida que aprendemos del uso inicial y de los comentarios (feedback).
---
### Uso más allá de los límites
Si también tienes créditos en tu saldo de Zen, puedes habilitar la opción **Use balance**
en la consola. Cuando está habilitada, Go recurrirá a tu saldo de Zen
después de que hayas alcanzado tus límites de uso en lugar de bloquear las peticiones.
---
### Por qué algunos modelos incluyen menos uso
Con Go, pagas $10/mes y nuestro objetivo es ofrecerte un uso equivalente a 6 veces esa cantidad.
Para la mayoría de los modelos, lo conseguimos mediante descuentos por volumen y capacidad de GPU reservada. Ese ahorro se traduce en un multiplicador de 6x para ti.
Para algunos modelos, todavía no hemos tenido la oportunidad de negociar un descuento o alojarlos a un coste menor, ya sea porque el modelo es nuevo o porque su precio público ya incluye un descuento.
Con estos modelos, aun así obtienes un poco más que si pagaras directamente a los proveedores de modelos; por eso su multiplicador de uso es menor en la tabla anterior.
---
## Endpoints
También puedes acceder a los modelos de Go a través de los siguientes endpoints de la API.
| Modelo | ID del modelo | Endpoint | Paquete de AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
El [ID del modelo](/docs/config/#models) en tu configuración de OpenCode
usa el formato `opencode-go/<model-id>`. Por ejemplo, para Kimi K3, usarías
`opencode-go/kimi-k3` en tu configuración.
---
### Modelos
Puedes obtener la lista completa de modelos disponibles y sus metadatos desde:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privacidad
El plan está diseñado principalmente para usuarios internacionales, con modelos alojados en US, EU y Singapore para ofrecer un acceso global estable. Nuestros proveedores siguen una política de retención cero y no utilizan tus datos para el entrenamiento de modelos.
---
## Objetivos
Creamos OpenCode Go para:
1. Hacer que la programación con IA sea **accesible** para más personas con una suscripción de bajo costo.
2. Proporcionar un acceso **confiable** a los mejores modelos abiertos de programación.
3. Seleccionar modelos que estén **probados y evaluados** para su uso en agentes de programación.
4. Evitar la **dependencia (lock-in)** al permitirte usar también cualquier otro proveedor con OpenCode.

224
packages/web/src/content/docs/fr/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: Abonnement à bas coût pour les modèles de codage ouverts.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go est un abonnement à bas coût — **5 $ pour votre premier mois**, puis **10 $/mois** — qui vous donne un accès fiable aux modèles de codage ouverts populaires.
Go fonctionne comme n'importe quel autre fournisseur dans OpenCode. Vous vous abonnez à OpenCode Go et obtenez votre clé d'API. C'est **totalement facultatif** et vous n'avez pas besoin de l'utiliser pour utiliser OpenCode.
Il est conçu principalement pour les utilisateurs internationaux, avec des modèles hébergés aux États-Unis, dans l'UE et à Singapour pour un accès mondial stable.
---
## Contexte
Les modèles ouverts sont devenus vraiment performants. Ils atteignent désormais des performances proches de celles des modèles propriétaires pour les tâches de codage. Et comme de nombreux fournisseurs peuvent les proposer de manière compétitive, ils sont généralement beaucoup moins chers.
Cependant, obtenir un accès fiable et à faible latence à ces modèles peut être difficile. Les fournisseurs varient en qualité et en disponibilité.
:::tip
Nous avons testé un groupe sélectionné de modèles et de fournisseurs qui fonctionnent bien avec OpenCode.
:::
Pour remédier à cela, nous avons fait plusieurs choses :
1. Nous avons testé un groupe sélectionné de modèles ouverts et discuté avec leurs équipes de la meilleure façon de les exécuter.
2. Nous avons ensuite travaillé avec quelques fournisseurs pour nous assurer qu'ils étaient correctement servis.
3. Enfin, nous avons évalué les performances de la combinaison modèle/fournisseur et avons dressé une liste que nous nous sentons à l'aise de recommander.
OpenCode Go vous donne accès à ces modèles pour **5 $ pour votre premier mois**, puis **10 $/mois**.
---
## Comment ça marche
OpenCode Go fonctionne comme n'importe quel autre fournisseur dans OpenCode.
1. Vous vous connectez à **<a href={console}>OpenCode Zen</a>**, vous vous abonnez à Go et copiez votre clé d'API.
2. Vous exécutez la commande `/connect` dans la TUI, sélectionnez `OpenCode Go` et collez votre clé d'API.
3. Exécutez `/models` dans la TUI pour voir la liste des modèles disponibles via Go.
:::note
Un seul membre par espace de travail peut s'abonner à OpenCode Go.
:::
La liste actuelle des modèles comprend :
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
La liste des modèles peut changer au fur et à mesure que nous en testons et en ajoutons de nouveaux.
---
## Limites d'utilisation
OpenCode Go inclut les limites suivantes :
- **Limite de 5 heures** — 12 $ d'utilisation
- **Limite hebdomadaire** — 30 $ d'utilisation
- **Limite mensuelle** — 60 $ d'utilisation
Les limites sont définies en valeur monétaire (dollars). Cela signifie que votre nombre réel de requêtes dépend du modèle que vous utilisez. Les modèles moins chers comme DeepSeek V4 Flash permettent plus de requêtes, tandis que les modèles plus coûteux comme GLM-5.2 en permettent moins.
Le tableau ci-dessous fournit une estimation du nombre de requêtes basée sur des modèles d'utilisation typiques de Go :
| Model | requêtes par 5 heures | requêtes par semaine | requêtes par mois |
| ----------------- | --------------------- | -------------------- | ----------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Les estimations sont basées sur les schémas de requêtes observés :
- Grok 4.5 — 1,100 tokens en entrée, 71,500 en cache, 220 tokens en sortie par requête
- GLM-5.2/5.1 — 700 tokens en entrée, 52,000 en cache, 150 tokens en sortie par requête
- Kimi K3 — 1,050 tokens en entrée, 76,500 en cache, 300 tokens en sortie par requête
- Kimi K2.7/K2.6 — 870 tokens en entrée, 55,000 en cache, 200 tokens en sortie par requête
- DeepSeek V4 Pro — 750 tokens en entrée, 82,000 en cache, 290 tokens en sortie par requête
- DeepSeek V4 Flash — 790 tokens en entrée, 68,000 en cache, 280 tokens en sortie par requête
- MiniMax M3 — 510 tokens en entrée, 56,000 en cache, 190 tokens en sortie par requête
- MiniMax M2.7 — 300 tokens en entrée, 55,000 en cache, 125 tokens en sortie par requête
- Qwen3.7 Max — 420 tokens en entrée, 66,000 en cache, 200 tokens en sortie par requête
- Qwen3.7 Plus — 500 tokens en entrée, 57,000 en cache, 190 tokens en sortie par requête
- Qwen3.6 Plus — 500 tokens en entrée, 57,000 en cache, 190 tokens en sortie par requête
- Hy3 — 830 tokens en entrée, 71,500 en cache, 295 tokens en sortie par requête
- MiMo-V2.5 — 830 tokens en entrée, 71,500 en cache, 295 tokens en sortie par requête
- MiMo-V2.5-Pro — 790 tokens en entrée, 86,000 en cache, 305 tokens en sortie par requête
Les estimations sont également basées sur les prix suivants par 1M tokens et sur l'utilisation mensuelle incluse avec chaque modèle :
| Modèle | Input | Output | Cached Read | Cached Write | Utilisation |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ----------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Vous pouvez suivre votre utilisation actuelle dans la **<a href={console}>console</a>**.
:::tip
Si vous atteignez la limite d'utilisation, vous pouvez continuer à utiliser les modèles gratuits.
:::
Les limites d'utilisation peuvent changer au fur et à mesure que nous tirons des enseignements des premières utilisations et des retours.
---
### Utilisation au-delà des limites
Si vous avez également des crédits sur votre solde Zen, vous pouvez activer l'option **Use balance** dans la console. Lorsqu'elle est activée, Go se rabattra sur votre solde Zen après que vous ayez atteint vos limites d'utilisation au lieu de bloquer les requêtes.
---
### Pourquoi certains modèles offrent un volume d'utilisation inférieur
Avec Go, vous payez 10 $/mois et nous cherchons à vous offrir un volume d'utilisation équivalant à 6 fois ce montant.
Pour la plupart des modèles, nous y parvenons grâce à des remises sur volume et à une capacité GPU réservée. Nous vous faisons ensuite bénéficier de ces économies grâce à un coefficient multiplicateur de 6.
Pour certains modèles, nous n'avons pas encore eu l'occasion de négocier une remise ou de les héberger à moindre coût, soit parce que le modèle est nouveau, soit parce que son tarif public est déjà réduit.
Pour ces modèles, vous obtenez tout de même un peu plus que si vous payiez directement les fournisseurs de modèles ; c'est pourquoi leur multiplicateur d'utilisation est plus faible dans le tableau ci-dessus.
---
## Points de terminaison
Vous pouvez également accéder aux modèles Go via les points de terminaison d'API suivants.
| Modèle | ID de modèle | Point de terminaison | Package AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
L'[ID de modèle](/docs/config/#models) dans votre configuration OpenCode utilise le format `opencode-go/<model-id>`. Par exemple, pour Kimi K3, vous utiliseriez `opencode-go/kimi-k3` dans votre configuration.
---
### Modèles
Vous pouvez récupérer la liste complète des modèles disponibles et leurs métadonnées à partir de :
```
https://opencode.ai/zen/go/v1/models
```
---
## Confidentialité
Cette offre est conçue avant tout pour les utilisateurs internationaux, avec des modèles hébergés aux US, dans l’EU et à Singapore afin d’assurer un accès mondial stable. Nos fournisseurs appliquent une politique de rétention zéro et n’utilisent pas vos données pour l’entraînement des modèles.
---
## Objectifs
Nous avons créé OpenCode Go pour :
1. Rendre le codage par IA **accessible** à un plus grand nombre de personnes avec un abonnement à bas coût.
2. Fournir un accès **fiable** aux meilleurs modèles de codage ouverts.
3. Sélectionner des modèles qui sont **testés et évalués** pour une utilisation en tant qu'agent de codage.
4. N'avoir **aucun verrouillage exclusif** en vous permettant d'utiliser n'importe quel autre fournisseur avec OpenCode également.

246
packages/web/src/content/docs/go.mdx
View file

@ -1,238 +1,38 @@
---
title: Go
description: Low cost subscription for open coding models.
description: Connect a named OpenCode Go subscription to OpenCode.
---
import config from "../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go is a low cost subscription that gives a named subscriber reliable access to popular open coding models in
OpenCode.
OpenCode Go is a low cost subscription — **$5 for your first month**, then **$10/month** — that gives you reliable access to popular open coding models.
Go works like any other provider in OpenCode. You subscribe to OpenCode Go and
get your API key. It's **completely optional** and you don't need to use it to
use OpenCode.
It is designed primarily for international users, with models hosted in the US, EU, and Singapore for stable global access.
---
## Background
Open models have gotten really good. They now reach performance close to
proprietary models for coding tasks. And because many providers can serve them
competitively, they are usually far cheaper.
However, getting reliable, low latency access to them can be difficult. Providers
vary in quality and availability.
:::tip
We tested a select group of models and providers that work well with OpenCode.
:::note
Go does not use a copied API key. Service accounts and other coding agents are not eligible.
:::
To fix this, we did a couple of things:
1. We tested a select group of open models and talked to their teams about how to
best run them.
2. We then worked with a few providers to make sure these were being served
correctly.
3. Finally, we benchmarked the combination of the model/provider and came up
with a list that we feel good recommending.
OpenCode Go gives you access to these models for **$5 for your first month**, then **$10/month**.
---
## How it works
OpenCode Go works like any other provider in OpenCode.
1. [Subscribe to Go in OpenCode Console](https://opencode.ai/console/go).
2. Run `opencode console login`.
3. Authorize the device in your browser.
4. Choose a model from the `opencode` provider in OpenCode.
1. You sign in to **<a href={console}>OpenCode Zen</a>**, subscribe to Go, and
copy your API key.
2. You run the `/connect` command in the TUI, select `OpenCode Go`, and paste
your API key.
3. Run `/models` in the TUI to see the list of models available through Go.
:::note
Only one member per workspace can subscribe to OpenCode Go.
:::
The current list of models includes:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
The list of models may change as we test and add new ones.
---
## Usage limits
OpenCode Go includes the following limits:
- **5 hour limit** — $12 of usage
- **Weekly limit** — $30 of usage
- **Monthly limit** — $60 of usage
Limits are defined in dollar value. This means your actual request count depends on the model you use. Cheaper models like DeepSeek V4 Flash allow for more requests, while higher-cost models like GLM-5.2 allow for fewer.
The table below provides an estimated request count based on typical Go usage patterns:
| Model | requests per 5 hour | requests per week | requests per month |
| ----------------- | ------------------- | ----------------- | ------------------ |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
The estimates are based on observed request patterns:
- Grok 4.5 — 1,100 input, 71,500 cached, 220 output tokens per request
- GLM-5.2/5.1 — 700 input, 52,000 cached, 150 output tokens per request
- Kimi K3 — 1,050 input, 76,500 cached, 300 output tokens per request
- Kimi K2.7/K2.6 — 870 input, 55,000 cached, 200 output tokens per request
- DeepSeek V4 Pro — 750 input, 82,000 cached, 290 output tokens per request
- DeepSeek V4 Flash — 790 input, 68,000 cached, 280 output tokens per request
- MiniMax M3 — 510 input, 56,000 cached, 190 output tokens per request
- MiniMax M2.7 — 300 input, 55,000 cached, 125 output tokens per request
- MiMo-V2.5 — 830 input, 71,500 cached, 295 output tokens per request
- MiMo-V2.5-Pro — 790 input, 86,000 cached, 305 output tokens per request
- Qwen3.7 Max — 420 input, 66,000 cached, 200 output tokens per request
- Qwen3.7 Plus — 500 input, 57,000 cached, 190 output tokens per request
- Qwen3.6 Plus — 500 input, 57,000 cached, 190 output tokens per request
- Hy3 — 830 input, 71,500 cached, 295 output tokens per request
The estimates are also based on the following prices per 1M tokens and the monthly usage included with each model:
| Model | Input | Output | Cached Read | Cached Write | Usage |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ----- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
You can track your current usage in the **<a href={console}>console</a>**.
:::tip
If you reach the usage limit, you can continue using the free models.
:::
Usage limits may change as we learn from early usage and feedback.
---
### Usage beyond limits
If you also have credits on your Zen balance, you can enable the **Use balance**
option in the console. When enabled, Go will fall back to your Zen balance
after you've reached your usage limits instead of blocking requests.
---
### Why some models have lower usage
With Go, you pay $10/month and we aim to give you 6x that in usage.
For most models, we make this work through bulk discounts and reserved GPU capacity. We then pass those savings on to you through the 6x multiplier.
For some models, we haven't had the opportunity to negotiate a discount or host them at a lower cost, either because the model is new or because their public pricing is already discounted.
For these models, you still get a little more than if you paid the model providers directly; this is why their usage mulitplier is lower in the table above.
---
## Endpoints
You can also access Go models through the following API endpoints.
| Model | Model ID | Endpoint | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
The [model id](/docs/config/#models) in your OpenCode config
uses the format `opencode-go/<model-id>`. For example, for Kimi K3, you would
use `opencode-go/kimi-k3` in your config.
---
### Models
You can fetch the full list of available models and their metadata from:
```
https://opencode.ai/zen/go/v1/models
```bash
opencode console login
```
---
For the initial launch, an account with exactly one organization connects automatically. Accounts with no organization
or multiple organizations stop with an actionable error instead of selecting an organization implicitly.
No separate Go provider or endpoint configuration is required. API keys and service-account credentials remain available
for ordinary OpenCode managed inference, but they do not grant Go subscription access.
## Pricing
Go costs **$5 for your first month**, then **$10/month**. Usage limits and the available model list are shown in
[OpenCode Console](https://opencode.ai/console/go).
## Privacy
The plan is designed primarily for international users, with models hosted in the US, EU, and Singapore for stable global access. Our providers follow a zero-retention policy and do not use your data for model training.
---
## Goals
We created OpenCode Go to:
1. Make AI coding **accessible** to more people with a low cost subscription.
2. Provide **reliable** access to the best open coding models.
3. Curate models that are **tested and benchmarked** for coding agent use.
4. Have **no lock-in** by allowing you to use any other provider with OpenCode as well.
Go is designed primarily for international users, with models hosted in the US, EU, and Singapore for stable global
access. Our providers follow a zero-retention policy and do not use your data for model training.

236
packages/web/src/content/docs/it/go.mdx
View file

@ -1,236 +0,0 @@
---
title: Go
description: Abbonamento a basso costo per modelli di programmazione aperti.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go è un abbonamento a basso costo — **5 $ per il primo mese**, poi **10 $/mese** — che ti offre un accesso affidabile ai popolari modelli di programmazione aperti.
Go funziona come qualsiasi altro provider in OpenCode. Ti abboni a OpenCode Go e
ottieni la tua chiave API. È **completamente facoltativo** e non hai bisogno di usarlo per
utilizzare OpenCode.
È progettato principalmente per gli utenti internazionali, con modelli ospitati negli Stati Uniti, nell'Unione Europea e a Singapore per un accesso globale stabile.
---
## Contesto
I modelli aperti sono diventati davvero validi. Ora raggiungono prestazioni vicine a quelle dei modelli proprietari per le attività di programmazione. E poiché molti provider possono fornirli in modo competitivo, di solito sono molto più economici.
Tuttavia, ottenere un accesso affidabile e a bassa latenza può essere difficile. I provider
variano per qualità e disponibilità.
:::tip
Abbiamo testato un gruppo selezionato di modelli e provider che funzionano bene con OpenCode.
:::
Per risolvere questo problema, abbiamo fatto un paio di cose:
1. Abbiamo testato un gruppo selezionato di modelli aperti e parlato con i loro team su come
eseguirli al meglio.
2. Abbiamo quindi collaborato con alcuni provider per assicurarci che venissero forniti
correttamente.
3. Infine, abbiamo eseguito dei benchmark sulla combinazione modello/provider e abbiamo stilato
un elenco che ci sentiamo di raccomandare.
OpenCode Go ti dà accesso a questi modelli a **5 $ per il primo mese**, poi a **10 $/mese**.
---
## Come funziona
OpenCode Go funziona come qualsiasi altro provider in OpenCode.
1. Accedi a **<a href={console}>OpenCode Zen</a>**, ti abboni a Go e
copi la tua chiave API.
2. Esegui il comando `/connect` nella TUI, selezioni `OpenCode Go` e incolli
la tua chiave API.
3. Esegui `/models` nella TUI per vedere l'elenco dei modelli disponibili tramite Go.
:::note
Solo un membro per workspace può abbonarsi a OpenCode Go.
:::
L'elenco attuale dei modelli include:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
L'elenco dei modelli potrebbe cambiare man mano che ne testiamo e aggiungiamo di nuovi.
---
## Limiti di utilizzo
OpenCode Go include i seguenti limiti:
- **Limite di 5 ore** — 12 $ di utilizzo
- **Limite settimanale** — 30 $ di utilizzo
- **Limite mensile** — 60 $ di utilizzo
I limiti sono definiti in valore in dollari. Questo significa che il conteggio effettivo delle richieste dipende dal modello utilizzato. Modelli più economici come DeepSeek V4 Flash consentono più richieste, mentre modelli più costosi come GLM-5.2 ne consentono di meno.
La tabella seguente fornisce una stima del conteggio delle richieste in base a pattern di utilizzo tipici di Go:
| Model | richieste ogni 5 ore | richieste a settimana | richieste al mese |
| ----------------- | -------------------- | --------------------- | ----------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Le stime si basano sui pattern di richieste osservati:
- Grok 4.5 — 1.100 di input, 71.500 in cache, 220 token di output per richiesta
- GLM-5.2/5.1 — 700 di input, 52.000 in cache, 150 token di output per richiesta
- Kimi K3 — 1.050 di input, 76.500 in cache, 300 token di output per richiesta
- Kimi K2.7/K2.6 — 870 di input, 55.000 in cache, 200 token di output per richiesta
- DeepSeek V4 Pro — 750 di input, 82.000 in cache, 290 token di output per richiesta
- DeepSeek V4 Flash — 790 di input, 68.000 in cache, 280 token di output per richiesta
- MiniMax M3 — 510 di input, 56.000 in cache, 190 token di output per richiesta
- MiniMax M2.7 — 300 di input, 55.000 in cache, 125 token di output per richiesta
- Qwen3.7 Max — 420 di input, 66.000 in cache, 200 token di output per richiesta
- Qwen3.7 Plus — 500 di input, 57.000 in cache, 190 token di output per richiesta
- Qwen3.6 Plus — 500 di input, 57.000 in cache, 190 token di output per richiesta
- Hy3 — 830 di input, 71.500 in cache, 295 token di output per richiesta
- MiMo-V2.5 — 830 di input, 71.500 in cache, 295 token di output per richiesta
- MiMo-V2.5-Pro — 790 di input, 86.000 in cache, 305 token di output per richiesta
Le stime si basano anche sui seguenti prezzi per 1M token e sull'utilizzo mensile incluso con ciascun modello:
| Modello | Input | Output | Cached Read | Cached Write | Utilizzo |
| ---------------------------- | ------ | ------ | ----------- | ------------ | -------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Puoi monitorare il tuo utilizzo attuale nella **<a href={console}>console</a>**.
:::tip
Se raggiungi il limite di utilizzo, puoi continuare a usare i modelli gratuiti.
:::
I limiti di utilizzo potrebbero cambiare man mano che raccogliamo dati sull'utilizzo iniziale e feedback.
---
### Utilizzo oltre i limiti
Se hai anche dei crediti nel tuo saldo Zen, puoi abilitare l'opzione **Use balance**
nella console. Quando è abilitata, Go ripiegherà sul tuo saldo Zen
dopo che avrai raggiunto i limiti di utilizzo invece di bloccare le richieste.
---
### Perché alcuni modelli hanno un utilizzo inferiore
Con Go, paghi $10/mese e puntiamo a offrirti un utilizzo pari a 6 volte tale importo.
Per la maggior parte dei modelli, ci riusciamo grazie a sconti sui volumi e capacità GPU riservata. Ti facciamo quindi beneficiare di questi risparmi tramite il moltiplicatore 6x.
Per alcuni modelli, non abbiamo ancora avuto l'opportunità di negoziare uno sconto o di ospitarli a un costo inferiore, perché il modello è nuovo oppure perché il prezzo pubblico è già scontato.
Per questi modelli, ottieni comunque un po' più di utilizzo rispetto a quanto otterresti pagando direttamente i provider dei modelli; per questo il loro moltiplicatore di utilizzo è più basso nella tabella precedente.
---
## Endpoint
Puoi anche accedere ai modelli Go tramite i seguenti endpoint API.
| Modello | ID Modello | Endpoint | Pacchetto AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
Il [model id](/docs/config/#models) nella tua OpenCode config
utilizza il formato `opencode-go/<model-id>`. Ad esempio, per Kimi K3, useresti
`opencode-go/kimi-k3` nella tua configurazione.
---
### Modelli
Puoi recuperare l'elenco completo dei modelli disponibili e i relativi metadati da:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privacy
Il piano è pensato principalmente per gli utenti internazionali, con modelli ospitati negli US, nell’EU e a Singapore per un accesso globale stabile. I nostri provider seguono una politica di zero-retention e non utilizzano i tuoi dati per l’addestramento dei modelli.
---
## Obiettivi
Abbiamo creato OpenCode Go per:
1. Rendere la programmazione con l'IA **accessibile** a più persone tramite un abbonamento a basso costo.
2. Fornire un accesso **affidabile** ai migliori modelli di programmazione aperti.
3. Curare modelli che sono **testati e valutati tramite benchmark** per l'uso con gli agenti di programmazione.
4. Non avere **nessun lock-in**, permettendoti di utilizzare anche qualsiasi altro provider con OpenCode.

224
packages/web/src/content/docs/ja/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: オープンなコーディングモデルを利用するための低価格サブスクリプション。
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Goは低価格のサブスクリプションで、**初月は5ドル**、その後は**月額10ドル**で、人気のオープンなコーディングモデルに安定してアクセスできます。
GoはOpenCodeの他のプロバイダーと同様に機能します。OpenCode GoをサブスクライブしてAPIキーを取得します。これは**完全に任意**であり、OpenCodeを使用するために必須ではありません。
主に海外ユーザー向けに設計されており、世界中で安定してアクセスできるよう、モデルは米国、EU、シンガポールでホストされています。
---
## 背景
オープンモデルは非常に優れています。現在では、コーディングタスクにおいてプロプライエタリ（独自）モデルに近いパフォーマンスを達成しています。そして、多くのプロバイダーが競争力を持って提供できるため、通常ははるかに安価です。
しかし、これらのモデルに安定して低遅延でアクセスすることは困難な場合があります。プロバイダーによって品質や可用性が異なります。
:::tip
OpenCodeでうまく動作する一部のモデルとプロバイダーをテストしました。
:::
この問題を解決するために、私たちはいくつかの取り組みを行いました：
1. 一部のオープンモデルをテストし、それらを最適に実行する方法について、そのチームと話し合いました。
2. 次に、これらが正しく提供されていることを確認するために、いくつかのプロバイダーと協力しました。
3. 最後に、モデルとプロバイダーの組み合わせをベンチマークし、自信を持ってお勧めできるリストを作成しました。
OpenCode Goを使用すると、これらのモデルに**初月は5ドル**、その後は**月額10ドル**でアクセスできます。
---
## 仕組み
OpenCode Goは、OpenCodeの他のプロバイダーと同様に機能します。
1. **<a href={console}>OpenCode Zen</a>**にサインインし、GoをサブスクライブしてAPIキーをコピーします。
2. TUIで`/connect`コマンドを実行し、`OpenCode Go`を選択して、APIキーを貼り付けます。
3. TUIで`/models`を実行すると、Goを通じて利用可能なモデルのリストが表示されます。
:::note
OpenCode Goをサブスクライブできるのは、1つのワークスペースにつき1メンバーのみです。
:::
現在のモデルリストには以下が含まれます：
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
新しいモデルをテストして追加するにつれて、モデルのリストは変更される場合があります。
---
## 利用制限
OpenCode Goには以下の制限が含まれています：
- **5時間の制限** — 12ドル分の利用
- **週間の制限** — 30ドル分の利用
- **月間の制限** — 60ドル分の利用
制限はドル単位で定義されています。つまり、実際のリクエスト数は使用するモデルによって異なります。DeepSeek V4 Flashのような安価なモデルではより多くのリクエストが可能ですが、GLM-5.2のような高コストのモデルではリクエスト数が少なくなります。
以下の表は、一般的なGoの利用パターンに基づいた推定リクエスト数を示しています：
| Model | 5時間あたりのリクエスト数 | 週間リクエスト数 | 月間リクエスト数 |
| ----------------- | ------------------------- | ---------------- | ---------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
推定値は、観測されたリクエストパターンに基づいています：
- Grok 4.5 — リクエストあたり 入力 1,100トークン、キャッシュ 71,500トークン、出力 220トークン
- GLM-5.2/5.1 — リクエストあたり 入力 700トークン、キャッシュ 52,000トークン、出力 150トークン
- Kimi K3 — リクエストあたり 入力 1,050トークン、キャッシュ 76,500トークン、出力 300トークン
- Kimi K2.7/K2.6 — リクエストあたり 入力 870トークン、キャッシュ 55,000トークン、出力 200トークン
- DeepSeek V4 Pro — リクエストあたり 入力 750トークン、キャッシュ 82,000トークン、出力 290トークン
- DeepSeek V4 Flash — リクエストあたり 入力 790トークン、キャッシュ 68,000トークン、出力 280トークン
- MiniMax M3 — リクエストあたり 入力 510トークン、キャッシュ 56,000トークン、出力 190トークン
- MiniMax M2.7 — リクエストあたり 入力 300トークン、キャッシュ 55,000トークン、出力 125トークン
- Qwen3.7 Max — リクエストあたり 入力 420トークン、キャッシュ 66,000トークン、出力 200トークン
- Qwen3.7 Plus — リクエストあたり 入力 500トークン、キャッシュ 57,000トークン、出力 190トークン
- Qwen3.6 Plus — リクエストあたり 入力 500トークン、キャッシュ 57,000トークン、出力 190トークン
- Hy3 — リクエストあたり 入力 830トークン、キャッシュ 71,500トークン、出力 295トークン
- MiMo-V2.5 — リクエストあたり 入力 830トークン、キャッシュ 71,500トークン、出力 295トークン
- MiMo-V2.5-Pro — リクエストあたり 入力 790トークン、キャッシュ 86,000トークン、出力 305トークン
推定値は、100万トークンあたりの以下の価格と、各モデルに含まれる月間利用枠にも基づいています：
| Model | Input | Output | Cached Read | Cached Write | Usage |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ----- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
現在の利用状況は**<a href={console}>コンソール</a>**で追跡できます。
:::tip
利用制限に達した場合でも、無料モデルを引き続き使用できます。
:::
初期の利用状況やフィードバックから学ぶにつれて、利用制限は変更される場合があります。
---
### 制限を超えた利用
Zen残高にクレジットがある場合は、コンソールで**Use balance**オプションを有効にすることができます。有効にすると、利用制限に達した後、リクエストをブロックする代わりに、GoはZen残高から利用するようになります。
---
### 一部のモデルの利用枠が少ない理由
Goでは月額$10を支払い、その6倍の利用枠を提供することを目指しています。
ほとんどのモデルでは、ボリュームディスカウントと予約済みのGPUキャパシティによってこれを実現しています。そして、その節約分を6倍の倍率で還元しています。
一部のモデルでは、モデルが新しい、または公開価格がすでに割引されているため、割引を交渉したり、より低コストでホストしたりする機会がまだありません。
これらのモデルでも、モデルプロバイダーに直接支払う場合より少し多く利用できます。そのため、上の表では利用枠の倍率が低くなっています。
---
## エンドポイント
以下のAPIエンドポイントを通じて、Goモデルにアクセスすることもできます。
| Model | Model ID | Endpoint | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
OpenCode設定の[model id](/docs/config/#models)は、`opencode-go/<model-id>`という形式を使用します。たとえば、Kimi K3の場合は、設定で`opencode-go/kimi-k3`を使用します。
---
### モデル
利用可能なモデルとそのメタデータの完全な一覧は、次から取得できます。
```
https://opencode.ai/zen/go/v1/models
```
---
## プライバシー
このプランは主に海外ユーザー向けに設計されており、安定したグローバルアクセスのため、モデルは US、EU、Singapore でホストされています。各プロバイダーはデータを保持しないポリシーに従っており、お客様のデータをモデルのトレーニングに使用することはありません。
---
## 目標
OpenCode Goを作成した目的は以下の通りです：
1. 低価格のサブスクリプションで、より多くの人がAIコーディングを**利用しやすく**する。
2. 最高のオープンなコーディングモデルに**安定した**アクセスを提供する。
3. コーディングエージェントの利用向けに**テストおよびベンチマークされた**モデルを厳選する。
4. OpenCodeで他のプロバイダーも使用できるようにすることで、**ベンダーロックインをなくす**。

224
packages/web/src/content/docs/ko/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: 오픈 코딩 모델을 위한 저비용 구독 서비스입니다.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go는 인기 있는 오픈 코딩 모델에 안정적으로 액세스할 수 있게 해주는 저비용 구독 서비스입니다. **첫 달은 $5**, 이후에는 **월 $10**입니다.
Go는 OpenCode의 다른 제공자와 똑같이 작동합니다. OpenCode Go를 구독하고 API 키를 발급받으면 됩니다. 이는 **완전히 선택 사항**이며, OpenCode를 사용하기 위해 꼭 필요하지는 않습니다.
주로 해외 사용자를 위해 설계되었으며, 안정적인 전 세계 액세스를 위해 모델은 미국, EU, 싱가포르에 호스팅됩니다.
---
## 배경
오픈 모델은 정말 많이 발전했습니다. 이제는 코딩 작업에서 독점 모델에 가까운 성능을 보여줍니다. 그리고 많은 제공자가 경쟁력 있게 서비스를 제공할 수 있기 때문에, 보통 훨씬 더 저렴합니다.
하지만 이런 모델에 안정적이고 낮은 지연 시간으로 액세스하는 일은 쉽지 않을 수 있습니다. 제공자마다 품질과 가용성이 다르기 때문입니다.
:::tip
OpenCode와 잘 맞는 선별된 모델과 제공자 그룹을 테스트했습니다.
:::
이를 해결하기 위해, 저희는 몇 가지를 진행했습니다.
1. 선별된 오픈 모델 그룹을 테스트하고, 각 팀과 가장 잘 운영하는 방법에 대해 논의했습니다.
2. 그런 다음 몇몇 제공자와 협력해, 이 모델들이 올바르게 서비스되도록 했습니다.
3. 마지막으로 모델/제공자 조합을 벤치마킹해, 자신 있게 추천할 수 있는 목록을 만들었습니다.
OpenCode Go를 사용하면 **첫 달은 $5**, 이후에는 **월 $10**으로 이러한 모델에 액세스할 수 있습니다.
---
## 작동 방식
OpenCode Go는 OpenCode의 다른 제공자와 똑같이 작동합니다.
1. **<a href={console}>OpenCode Zen</a>**에 로그인해 Go를 구독하고 API 키를 복사합니다.
2. TUI에서 `/connect` 명령을 실행하고 `OpenCode Go`를 선택한 다음 API 키를 붙여넣습니다.
3. TUI에서 `/models`를 실행해 Go를 통해 사용할 수 있는 모델 목록을 확인합니다.
:::note
workspace당 한 명의 멤버만 OpenCode Go를 구독할 수 있습니다.
:::
현재 모델 목록에는 다음이 포함됩니다.
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
새로운 모델을 테스트하고 추가함에 따라 이 목록은 변경될 수 있습니다.
---
## 사용 한도
OpenCode Go에는 다음과 같은 한도가 포함됩니다.
- **5시간 한도** — 사용량 $12
- **주간 한도** — 사용량 $30
- **월간 한도** — 사용량 $60
한도는 달러 금액 기준으로 정의됩니다. 즉, 실제 요청 횟수는 사용하는 모델에 따라 달라집니다. DeepSeek V4 Flash처럼 저렴한 모델은 더 많은 요청이 가능하고, GLM-5.2처럼 비용이 더 높은 모델은 더 적은 요청이 가능합니다.
아래 표는 일반적인 Go 사용 패턴을 기준으로 한 예상 요청 횟수를 보여줍니다.
| Model | 5시간당 요청 횟수 | 주간 요청 횟수 | 월간 요청 횟수 |
| ----------------- | ----------------- | -------------- | -------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
이 예상치는 관찰된 요청 패턴을 기준으로 합니다.
- Grok 4.5 — 요청당 입력 1,100, 캐시 71,500, 출력 토큰 220
- GLM-5.2/5.1 — 요청당 입력 700, 캐시 52,000, 출력 토큰 150
- Kimi K3 — 요청당 입력 1,050, 캐시 76,500, 출력 토큰 300
- Kimi K2.7/K2.6 — 요청당 입력 870, 캐시 55,000, 출력 토큰 200
- DeepSeek V4 Pro — 요청당 입력 750, 캐시 82,000, 출력 토큰 290
- DeepSeek V4 Flash — 요청당 입력 790, 캐시 68,000, 출력 토큰 280
- MiniMax M3 — 요청당 입력 510, 캐시 56,000, 출력 토큰 190
- MiniMax M2.7 — 요청당 입력 300, 캐시 55,000, 출력 토큰 125
- Qwen3.7 Max — 요청당 입력 420, 캐시 66,000, 출력 토큰 200
- Qwen3.7 Plus — 요청당 입력 500, 캐시 57,000, 출력 토큰 190
- Qwen3.6 Plus — 요청당 입력 500, 캐시 57,000, 출력 토큰 190
- Hy3 — 요청당 입력 830, 캐시 71,500, 출력 토큰 295
- MiMo-V2.5 — 요청당 입력 830, 캐시 71,500, 출력 토큰 295
- MiMo-V2.5-Pro — 요청당 입력 790, 캐시 86,000, 출력 토큰 305
이 예상치는 또한 1M tokens당 다음 가격과 각 모델에 포함된 월간 사용량을 기준으로 합니다.
| Model | Input | Output | Cached Read | Cached Write | Usage |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ----- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
현재 사용량은 **<a href={console}>console</a>**에서 확인할 수 있습니다.
:::tip
사용 한도에 도달해도 무료 모델은 계속 사용할 수 있습니다.
:::
초기 사용량과 피드백을 바탕으로 사용 한도는 변경될 수 있습니다.
---
### 한도 초과 사용
Zen 잔액에 크레딧도 있다면, console에서 **Use balance** 옵션을 활성화할 수 있습니다. 이 옵션을 켜면, 사용 한도에 도달했을 때 요청을 차단하는 대신 Go가 Zen 잔액으로 자동 전환됩니다.
---
### 일부 모델의 사용량이 더 적은 이유
Go에서는 월 $10를 지불하며, 저희는 그 6배의 사용량을 제공하는 것을 목표로 합니다.
대부분의 모델은 대량 할인과 예약된 GPU 용량을 통해 이를 실현합니다. 그런 다음 6배의 사용량 배율을 통해 절감 혜택을 사용자에게 돌려드립니다.
일부 모델은 새로 출시되었거나 공개 가격에 이미 할인이 적용되어 있어, 아직 할인을 협상하거나 더 저렴한 비용으로 호스팅할 기회가 없었습니다.
이러한 모델도 모델 제공자에게 직접 비용을 지불할 때보다 약간 더 많은 사용량을 제공합니다. 이것이 위 표에서 해당 모델의 사용량 배율이 더 낮은 이유입니다.
---
## 엔드포인트
다음 API 엔드포인트를 통해서도 Go 모델에 액세스할 수 있습니다.
| 모델 | 모델 ID | 엔드포인트 | AI SDK 패키지 |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
OpenCode config의 [model id](/docs/config/#models)는 `opencode-go/<model-id>` 형식을 사용합니다. 예를 들어 Kimi K3의 경우 config에서 `opencode-go/kimi-k3`를 사용하면 됩니다.
---
### 모델
사용 가능한 전체 모델 목록과 메타데이터는 다음에서 가져올 수 있습니다.
```
https://opencode.ai/zen/go/v1/models
```
---
## 개인정보 보호
이 플랜은 안정적인 전 세계 액세스를 위해 모델을 미국, EU, 싱가포르에 호스팅하며, 주로 해외 사용자를 위해 설계되었습니다. 저희 제공자는 zero-retention 정책을 따르며, 고객 데이터를 모델 학습에 사용하지 않습니다.
---
## 목표
OpenCode Go를 만든 목적은 다음과 같습니다.
1. 저비용 구독으로 더 많은 사람이 AI 코딩을 **접근 가능하게** 이용할 수 있도록 합니다.
2. 최고의 오픈 코딩 모델에 **안정적으로** 액세스할 수 있게 합니다.
3. 코딩 에이전트 사용에 맞게 **테스트되고 벤치마킹된** 모델을 선별합니다.
4. OpenCode에서 다른 제공자도 함께 사용할 수 있게 해 **종속성 없음**을 유지합니다.

238
packages/web/src/content/docs/nb/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Lavkostnadsabonnement for åpne kodemodeller.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go er et lavkostnadsabonnement — **$5 for din første måned**, deretter **$10/måned** — som gir deg pålitelig tilgang til populære åpne kodemodeller.
Go fungerer som enhver annen leverandør i OpenCode. Du abonnerer på OpenCode Go og
får din API-nøkkel. Det er **helt valgfritt**, og du trenger ikke å bruke det for å
bruke OpenCode.
Det er primært utformet for internasjonale brukere, med modeller driftet i USA, EU og Singapore for stabil global tilgang.
---
## Bakgrunn
Åpne modeller har blitt veldig gode. De oppnår nå ytelse nær
proprietære modeller for kodeoppgaver. Og fordi mange leverandører kan tilby dem
konkurransedyktig, er de vanligvis mye billigere.
Imidlertid kan det være vanskelig å få pålitelig tilgang med lav forsinkelse til dem. Leverandører
varierer i kvalitet og tilgjengelighet.
:::tip
Vi testet en utvalgt gruppe modeller og leverandører som fungerer godt med OpenCode.
:::
For å fikse dette, gjorde vi et par ting:
1. Vi testet en utvalgt gruppe åpne modeller og snakket med teamene deres om hvordan de
best kan kjøres.
2. Deretter samarbeidet vi med noen få leverandører for å sikre at disse ble levert
riktig.
3. Til slutt utførte vi ytelsestester på kombinasjonen av modell og leverandør, og kom frem
til en liste som vi trygt kan anbefale.
OpenCode Go gir deg tilgang til disse modellene for **$5 for din første måned**, deretter **$10/måned**.
---
## Hvordan det fungerer
OpenCode Go fungerer som enhver annen leverandør i OpenCode.
1. Du logger inn på **<a href={console}>OpenCode Zen</a>**, abonnerer på Go, og
kopierer din API-nøkkel.
2. Du kjører kommandoen `/connect` i TUI-en, velger `OpenCode Go`, og limer inn
din API-nøkkel.
3. Kjør `/models` i TUI-en for å se listen over modeller som er tilgjengelige gjennom Go.
:::note
Kun ett medlem per arbeidsområde kan abonnere på OpenCode Go.
:::
Den nåværende listen over modeller inkluderer:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Listen over modeller kan endres etter hvert som vi tester og legger til nye.
---
## Bruksgrenser
OpenCode Go inkluderer følgende grenser:
- **5-timers grense** — $12 i bruk
- **Ukentlig grense** — $30 i bruk
- **Månedlig grense** — $60 i bruk
Grensene er definert i dollarverdi. Dette betyr at ditt faktiske antall forespørsler avhenger av modellen du bruker. Billigere modeller som DeepSeek V4 Flash tillater flere forespørsler, mens dyrere modeller som GLM-5.2 tillater færre.
Tabellen nedenfor gir et estimert antall forespørsler basert på typiske bruksmønstre for Go:
| Model | forespørsler per 5 timer | forespørsler per uke | forespørsler per måned |
| ----------------- | ------------------------ | -------------------- | ---------------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Estimatene er basert på observerte forespørselsmønstre:
- Grok 4.5 — 1 100 input, 71 500 bufret, 220 output-tokens per forespørsel
- GLM-5.2/5.1 — 700 input, 52 000 bufret, 150 output-tokens per forespørsel
- Kimi K3 — 1 050 input, 76 500 bufret, 300 output-tokens per forespørsel
- Kimi K2.7/K2.6 — 870 input, 55 000 bufret, 200 output-tokens per forespørsel
- DeepSeek V4 Pro — 750 input, 82 000 bufret, 290 output-tokens per forespørsel
- DeepSeek V4 Flash — 790 input, 68 000 bufret, 280 output-tokens per forespørsel
- MiniMax M3 — 510 input, 56 000 bufret, 190 output-tokens per forespørsel
- MiniMax M2.7 — 300 input, 55 000 bufret, 125 output-tokens per forespørsel
- Qwen3.7 Max — 420 input, 66 000 bufret, 200 output-tokens per forespørsel
- Qwen3.7 Plus — 500 input, 57 000 bufret, 190 output-tokens per forespørsel
- Qwen3.6 Plus — 500 input, 57 000 bufret, 190 output-tokens per forespørsel
- Hy3 — 830 input, 71 500 bufret, 295 output-tokens per forespørsel
- MiMo-V2.5 — 830 input, 71 500 bufret, 295 output-tokens per forespørsel
- MiMo-V2.5-Pro — 790 input, 86 000 bufret, 305 output-tokens per forespørsel
Estimatene er også basert på følgende priser per 1M tokens og den månedlige bruken som er inkludert med hver modell:
| Model | Input | Output | Cached Read | Cached Write | Bruk |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ---- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Du kan spore din nåværende bruk i **<a href={console}>konsollen</a>**.
:::tip
Hvis du når bruksgrensen, kan du fortsette å bruke gratismodellene.
:::
Bruksgrenser kan endres etter hvert som vi lærer fra tidlig bruk og tilbakemeldinger.
---
### Bruk utover grensene
Hvis du også har kreditt på din Zen-saldo, kan du aktivere alternativet **Bruk saldo**
i konsollen. Når dette er aktivert, vil Go falle tilbake på din Zen-saldo
etter at du har nådd bruksgrensene dine, i stedet for å blokkere forespørsler.
---
### Hvorfor noen modeller har lavere bruk
Med Go betaler du $10/måned, og vi har som mål å gi deg seks ganger så mye bruk.
For de fleste modeller får vi dette til gjennom volumrabatter og reservert GPU-kapasitet. Deretter gir vi disse besparelsene videre til deg gjennom 6x-multiplikatoren.
For noen modeller har vi ikke hatt muligheten til å forhandle frem en rabatt eller drifte dem til en lavere kostnad, enten fordi modellen er ny, eller fordi den offentlige prisen allerede er rabattert.
For disse modellene får du fortsatt litt mer enn om du betalte modellleverandørene direkte. Derfor er bruksmultiplikatoren deres lavere i tabellen ovenfor.
---
## Endepunkter
Du kan også få tilgang til Go-modeller gjennom følgende API-endepunkter.
| Modell | Modell-ID | Endepunkt | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
[Modell-ID-en](/docs/config/#models) i din OpenCode-konfigurasjon
bruker formatet `opencode-go/<model-id>`. For eksempel, for Kimi K3, vil du
bruke `opencode-go/kimi-k3` i konfigurasjonen din.
---
### Modeller
Du kan hente hele listen over tilgjengelige modeller og metadataene deres fra:
```
https://opencode.ai/zen/go/v1/models
```
---
## Personvern
Planen er primært utformet for internasjonale brukere, med modeller hostet i US, EU og Singapore for stabil global tilgang. Våre leverandører følger en zero-retention-policy og bruker ikke dataene dine til modelltrening.
---
## Mål
Vi opprettet OpenCode Go for å:
1. Gjøre AI-koding **tilgjengelig** for flere mennesker med et lavkostnadsabonnement.
2. Gi **pålitelig** tilgang til de beste åpne kodemodellene.
3. Kuratere modeller som er **testet og ytelsesmålt** for bruk av kodeagenter.
4. Ha **ingen innelåsing** ved å la deg bruke en hvilken som helst annen leverandør med OpenCode også.

230
packages/web/src/content/docs/pl/go.mdx
View file

@ -1,230 +0,0 @@
---
title: Go
description: Niskokosztowa subskrypcja na otwarte modele do kodowania.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go to niskokosztowa subskrypcja — **5 $ za pierwszy miesiąc**, a następnie **10 $/miesiąc** — która zapewnia niezawodny dostęp do popularnych otwartych modeli do kodowania.
Go działa jak każdy inny dostawca w OpenCode. Subskrybujesz OpenCode Go i
otrzymujesz swój klucz API. Jest to **całkowicie opcjonalne** i nie musisz z tego korzystać, aby
używać OpenCode.
Jest przeznaczony przede wszystkim dla użytkowników międzynarodowych, a modele są hostowane w USA, UE i Singapurze, co zapewnia stabilny globalny dostęp.
---
## Tło
Otwarte modele stały się naprawdę dobre. Obecnie osiągają wydajność zbliżoną do modeli własnościowych w zadaniach związanych z kodowaniem. A ponieważ wielu dostawców może je oferować konkurencyjnie, są one zazwyczaj znacznie tańsze.
Jednak uzyskanie do nich niezawodnego dostępu o niskich opóźnieniach może być trudne. Dostawcy różnią się pod względem jakości i dostępności.
:::tip
Przetestowaliśmy wybraną grupę modeli i dostawców, którzy dobrze współpracują z OpenCode.
:::
Aby to naprawić, zrobiliśmy kilka rzeczy:
1. Przetestowaliśmy wybraną grupę otwartych modeli i porozmawialiśmy z ich zespołami na temat najlepszych sposobów ich uruchamiania.
2. Następnie nawiązaliśmy współpracę z kilkoma dostawcami, aby upewnić się, że są one obsługiwane poprawnie.
3. Na koniec przetestowaliśmy kombinację modelu/dostawcy i stworzyliśmy listę, którą możemy z przekonaniem polecić.
OpenCode Go daje Ci dostęp do tych modeli za **5 $ za pierwszy miesiąc**, a następnie **10 $/miesiąc**.
---
## Jak to działa
OpenCode Go działa jak każdy inny dostawca w OpenCode.
1. Logujesz się do **<a href={console}>OpenCode Zen</a>**, subskrybujesz Go i
kopiujesz swój klucz API.
2. Uruchamiasz komendę `/connect` w TUI, wybierasz `OpenCode Go` i wklejasz
swój klucz API.
3. Uruchom `/models` w TUI, aby zobaczyć listę modeli dostępnych w ramach Go.
:::note
Tylko jeden członek na obszar roboczy (workspace) może zasubskrybować OpenCode Go.
:::
Obecna lista modeli obejmuje:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Lista modeli może ulec zmianie w miarę testowania i dodawania nowych.
---
## Limity użycia
OpenCode Go zawiera następujące limity:
- **Limit 5-godzinny** — użycie o wartości 12 $
- **Limit tygodniowy** — użycie o wartości 30 $
- **Limit miesięczny** — użycie o wartości 60 $
Limity są zdefiniowane w wartości w dolarach. Oznacza to, że rzeczywista liczba żądań zależy od używanego modelu. Tańsze modele, takie jak DeepSeek V4 Flash, pozwalają na więcej żądań, podczas gdy modele o wyższym koszcie, takie jak GLM-5.2, pozwalają na mniej.
Poniższa tabela przedstawia szacunkową liczbę żądań na podstawie typowych wzorców korzystania z Go:
| Model | żądania na 5 godzin | żądania na tydzień | żądania na miesiąc |
| ----------------- | ------------------- | ------------------ | ------------------ |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Szacunki te opierają się na zaobserwowanych wzorcach żądań:
- Grok 4.5 — 1 100 tokenów wejściowych, 71 500 w pamięci podręcznej, 220 tokenów wyjściowych na żądanie
- GLM-5.2/5.1 — 700 tokenów wejściowych, 52 000 w pamięci podręcznej, 150 tokenów wyjściowych na żądanie
- Kimi K3 — 1 050 tokenów wejściowych, 76 500 w pamięci podręcznej, 300 tokenów wyjściowych na żądanie
- Kimi K2.7/K2.6 — 870 tokenów wejściowych, 55 000 w pamięci podręcznej, 200 tokenów wyjściowych na żądanie
- DeepSeek V4 Pro — 750 tokenów wejściowych, 82 000 w pamięci podręcznej, 290 tokenów wyjściowych na żądanie
- DeepSeek V4 Flash — 790 tokenów wejściowych, 68 000 w pamięci podręcznej, 280 tokenów wyjściowych na żądanie
- MiniMax M3 — 510 tokenów wejściowych, 56 000 w pamięci podręcznej, 190 tokenów wyjściowych na żądanie
- MiniMax M2.7 — 300 tokenów wejściowych, 55 000 w pamięci podręcznej, 125 tokenów wyjściowych na żądanie
- Qwen3.7 Max — 420 tokenów wejściowych, 66 000 w pamięci podręcznej, 200 tokenów wyjściowych na żądanie
- Qwen3.7 Plus — 500 tokenów wejściowych, 57 000 w pamięci podręcznej, 190 tokenów wyjściowych na żądanie
- Qwen3.6 Plus — 500 tokenów wejściowych, 57 000 w pamięci podręcznej, 190 tokenów wyjściowych na żądanie
- Hy3 — 830 tokenów wejściowych, 71 500 w pamięci podręcznej, 295 tokenów wyjściowych na żądanie
- MiMo-V2.5 — 830 tokenów wejściowych, 71 500 w pamięci podręcznej, 295 tokenów wyjściowych na żądanie
- MiMo-V2.5-Pro — 790 tokenów wejściowych, 86 000 w pamięci podręcznej, 305 tokenów wyjściowych na żądanie
Szacunki opierają się również na następujących cenach za 1M tokenów oraz miesięcznym użyciu dostępnym dla każdego modelu:
| Model | Wejście | Wyjście | Odczyt z cache | Zapis do cache | Użycie |
| ---------------------------- | ------- | ------- | -------------- | -------------- | ------ |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Możesz śledzić swoje bieżące zużycie w **<a href={console}>konsoli</a>**.
:::tip
Jeśli osiągniesz limit użycia, możesz nadal korzystać z darmowych modeli.
:::
Limity użycia mogą ulec zmianie w miarę wyciągania wniosków z wczesnego użytkowania i opinii.
---
### Użycie ponad limity
Jeśli masz również środki na swoim saldzie Zen, możesz włączyć opcję **Użyj salda** (Use balance) w konsoli. Po włączeniu Go będzie korzystać z Twojego salda Zen po osiągnięciu limitów użycia zamiast blokować żądania.
---
### Dlaczego niektóre modele mają niższe limity użycia
Go kosztuje $10/miesiąc, a naszym celem jest zapewnienie Ci użycia o wartości 6x większej niż ta kwota.
W przypadku większości modeli jest to możliwe dzięki rabatom hurtowym i zarezerwowanej mocy obliczeniowej GPU. Uzyskane w ten sposób oszczędności przekazujemy Tobie w postaci mnożnika 6x.
W przypadku niektórych modeli nie mieliśmy jeszcze możliwości wynegocjowania rabatu ani hostowania ich po niższym koszcie, ponieważ model jest nowy lub jego publiczny cennik już uwzględnia rabat.
W przypadku tych modeli nadal otrzymujesz nieco więcej, niż płacąc bezpośrednio ich dostawcom. Dlatego ich mnożnik użycia w powyższej tabeli jest niższy.
---
## Punkty końcowe
Możesz również uzyskać dostęp do modeli Go za pośrednictwem następujących punktów końcowych API.
| Model | ID modelu | Punkt końcowy | Pakiet AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
[ID modelu](/docs/config/#models) w Twojej konfiguracji OpenCode
używa formatu `opencode-go/<model-id>`. Na przykład dla Kimi K3 należy użyć
`opencode-go/kimi-k3` w swojej konfiguracji.
---
### Modele
Pełną listę dostępnych modeli i ich metadane możesz pobrać z:
```
https://opencode.ai/zen/go/v1/models
```
---
## Prywatność
Plan został zaprojektowany przede wszystkim z myślą o użytkownikach międzynarodowych, a modele są hostowane w US, EU i Singapore, aby zapewnić stabilny dostęp na całym świecie. Nasi dostawcy stosują politykę zerowej retencji i nie wykorzystują Twoich danych do trenowania modeli.
---
## Cele
Stworzyliśmy OpenCode Go, aby:
1. Uczynić kodowanie z pomocą sztucznej inteligencji **dostępnym** dla większej liczby osób dzięki niskokosztowej subskrypcji.
2. Zapewnić **niezawodny** dostęp do najlepszych otwartych modeli do kodowania.
3. Wybierać modele, które są **przetestowane i sprawdzone w testach wydajności** do użytku przez agentów kodujących.
4. Zapewnić **brak ograniczeń (no lock-in)**, pozwalając na korzystanie z dowolnego innego dostawcy wraz z OpenCode.

28
packages/web/src/content/docs/providers.mdx
View file

@ -125,30 +125,18 @@ It works like any other provider in OpenCode and is completely optional to use.
OpenCode Go is a low cost subscription plan that provides reliable access to popular open coding models provided by the OpenCode team that have been
tested and verified to work well with OpenCode.
1. Run the `/connect` command in the TUI, select `OpenCode Go`, and head to [opencode.ai/auth](https://opencode.ai/zen).
1. [Subscribe to Go in OpenCode Console](https://opencode.ai/console/go).
```txt
/connect
2. Run the Console device login and authorize it in your browser.
```bash
opencode console login
```
2. Sign in, add your billing details, and copy your API key.
3. Choose a model from the `opencode` provider.
3. Paste your API key.
```txt
┌ API key
│
│
└ enter
```
4. Run `/models` in the TUI to see the list of models we recommend.
```txt
/models
```
It works like any other provider in OpenCode and is completely optional to use.
Go belongs to the named subscriber and works only through OpenCode's `opencode` provider. Do not copy an API key.
Service accounts can use ordinary managed inference, but they are not eligible for Go.
---

238
packages/web/src/content/docs/pt-br/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Assinatura de baixo custo para modelos abertos de programação.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
O OpenCode Go é uma assinatura de baixo custo — **US$ 5 no seu primeiro mês**, depois **US$ 10/mês** — que oferece acesso confiável a modelos abertos de programação populares.
O Go funciona como qualquer outro provedor no OpenCode. Você assina o OpenCode Go e
obtém a sua chave de API. Ele é **totalmente opcional** e você não precisa usá-lo para
usar o OpenCode.
Ele foi projetado principalmente para usuários internacionais, com modelos hospedados nos EUA, na UE e em Singapura para um acesso global estável.
---
## Contexto
Os modelos abertos ficaram muito bons. Eles agora alcançam um desempenho próximo aos
modelos proprietários para tarefas de programação. E como muitos provedores podem servi-los
de forma competitiva, eles geralmente são muito mais baratos.
No entanto, obter um acesso confiável e de baixa latência a eles pode ser difícil. Os provedores
variam em qualidade e disponibilidade.
:::tip
Testamos um grupo selecionado de modelos e provedores que funcionam bem com o OpenCode.
:::
Para resolver isso, fizemos algumas coisas:
1. Testamos um grupo selecionado de modelos abertos e conversamos com suas equipes sobre a
melhor forma de executá-los.
2. Em seguida, trabalhamos com alguns provedores para garantir que eles estivessem sendo servidos
corretamente.
3. Por fim, avaliamos por benchmark a combinação de modelo/provedor e elaboramos
uma lista que nos sentimos confortáveis em recomendar.
O OpenCode Go lhe dá acesso a esses modelos por **US$ 5 no seu primeiro mês**, depois **US$ 10/mês**.
---
## Como funciona
O OpenCode Go funciona como qualquer outro provedor no OpenCode.
1. Você entra no **<a href={console}>OpenCode Zen</a>**, assina o Go e
copia a sua chave de API.
2. Você executa o comando `/connect` na TUI, seleciona `OpenCode Go` e cola
a sua chave de API.
3. Execute `/models` na TUI para ver a lista de modelos disponíveis através do Go.
:::note
Apenas um membro por workspace pode assinar o OpenCode Go.
:::
A lista atual de modelos inclui:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
A lista de modelos pode mudar conforme testamos e adicionamos novos.
---
## Limites de uso
O OpenCode Go inclui os seguintes limites:
- **Limite de 5 horas** — US$ 12 de uso
- **Limite semanal** — US$ 30 de uso
- **Limite mensal** — US$ 60 de uso
Os limites são definidos em valor em dólares. Isso significa que a sua contagem real de requisições depende do modelo que você usa. Modelos mais baratos como o DeepSeek V4 Flash permitem mais requisições, enquanto modelos de custo mais alto como o GLM-5.2 permitem menos.
A tabela abaixo fornece uma contagem estimada de requisições com base nos padrões típicos de uso do Go:
| Model | requisições por 5 horas | requisições por semana | requisições por mês |
| ----------------- | ----------------------- | ---------------------- | ------------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
As estimativas se baseiam nos padrões de requisições observados:
- Grok 4.5 — 1.100 tokens de entrada, 71.500 em cache, 220 tokens de saída por requisição
- GLM-5.2/5.1 — 700 tokens de entrada, 52.000 em cache, 150 tokens de saída por requisição
- Kimi K3 — 1.050 tokens de entrada, 76.500 em cache, 300 tokens de saída por requisição
- Kimi K2.7/K2.6 — 870 tokens de entrada, 55.000 em cache, 200 tokens de saída por requisição
- DeepSeek V4 Pro — 750 tokens de entrada, 82.000 em cache, 290 tokens de saída por requisição
- DeepSeek V4 Flash — 790 tokens de entrada, 68.000 em cache, 280 tokens de saída por requisição
- MiniMax M3 — 510 tokens de entrada, 56.000 em cache, 190 tokens de saída por requisição
- MiniMax M2.7 — 300 tokens de entrada, 55.000 em cache, 125 tokens de saída por requisição
- Qwen3.7 Max — 420 tokens de entrada, 66.000 em cache, 200 tokens de saída por requisição
- Qwen3.7 Plus — 500 tokens de entrada, 57.000 em cache, 190 tokens de saída por requisição
- Qwen3.6 Plus — 500 tokens de entrada, 57.000 em cache, 190 tokens de saída por requisição
- Hy3 — 830 tokens de entrada, 71.500 em cache, 295 tokens de saída por requisição
- MiMo-V2.5 — 830 tokens de entrada, 71.500 em cache, 295 tokens de saída por requisição
- MiMo-V2.5-Pro — 790 tokens de entrada, 86.000 em cache, 305 tokens de saída por requisição
As estimativas também se baseiam nos seguintes preços por 1M tokens e no uso mensal incluído com cada modelo:
| Modelo | Entrada | Saída | Leitura em cache | Escrita em cache | Uso |
| ---------------------------- | ------- | ------ | ---------------- | ---------------- | --- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Você pode acompanhar o seu uso atual no **<a href={console}>console</a>**.
:::tip
Se você atingir o limite de uso, pode continuar usando os modelos gratuitos.
:::
Os limites de uso podem mudar à medida que aprendemos com o uso inicial e o feedback.
---
### Uso além dos limites
Se você também tiver créditos no seu saldo do Zen, pode habilitar a opção **Use balance**
no console. Quando habilitada, o Go usará o seu saldo do Zen como alternativa
após você atingir os seus limites de uso em vez de bloquear as requisições.
---
### Por que alguns modelos têm um uso menor
Com o Go, você paga $10/mês, e nosso objetivo é oferecer 6x esse valor em uso.
Para a maioria dos modelos, conseguimos fazer isso por meio de descontos por volume e capacidade reservada de GPU. Repassamos essa economia a você por meio do multiplicador de 6x.
Para alguns modelos, ainda não tivemos a oportunidade de negociar um desconto ou hospedá-los a um custo menor, seja porque o modelo é novo ou porque o preço público já tem desconto.
Para esses modelos, você ainda recebe um pouco mais do que receberia se pagasse diretamente aos provedores dos modelos; por isso, o multiplicador de uso deles é menor na tabela acima.
---
## Endpoints
Você também pode acessar os modelos do Go através dos seguintes endpoints de API.
| Modelo | ID do Modelo | Endpoint | Pacote do AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
O [ID do modelo](/docs/config/#models) na sua configuração do OpenCode
usa o formato `opencode-go/<model-id>`. Por exemplo, para o Kimi K3, você usaria
`opencode-go/kimi-k3` na sua configuração.
---
### Modelos
Você pode buscar a lista completa de modelos disponíveis e seus metadados em:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privacidade
O plano foi projetado principalmente para usuários internacionais, com modelos hospedados em US, EU e Singapore para garantir acesso global estável. Nossos provedores seguem uma política de retenção zero e não usam seus dados para treinamento de modelos.
---
## Objetivos
Criamos o OpenCode Go para:
1. Tornar a programação com IA **acessível** a mais pessoas com uma assinatura de baixo custo.
2. Fornecer acesso **confiável** aos melhores modelos abertos de programação.
3. Ter modelos sob curadoria que são **testados e avaliados (benchmarked)** para o uso por agentes de programação.
4. Não ter **nenhuma restrição de fornecedor (lock-in)**, permitindo que você use qualquer outro provedor com o OpenCode também.

238
packages/web/src/content/docs/ru/go.mdx
View file

@ -1,238 +0,0 @@
---
title: Go
description: Недорогая подписка на открытые модели для программирования.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go — это недорогая подписка (**$5 за первый месяц**, далее **$10 в месяц**), которая предоставляет надежный доступ к популярным открытым моделям для программирования.
Go работает так же, как и любой другой провайдер в OpenCode. Вы оформляете подписку на OpenCode Go и
получаете свой API-ключ. Использование Go **абсолютно необязательно**, и вам не нужно использовать его, чтобы
пользоваться OpenCode.
Она предназначена в первую очередь для пользователей со всего мира, а модели размещены в США, ЕС и Сингапуре для стабильного глобального доступа.
---
## Предпосылки
Открытые модели стали действительно хороши. Сейчас они достигают производительности, близкой к
проприетарным моделям для задач программирования. И поскольку многие провайдеры могут предоставлять к ним доступ
на конкурентных условиях, они, как правило, обходятся гораздо дешевле.
Однако получить к ним надежный доступ с низкой задержкой может быть непросто. Провайдеры
различаются по качеству и доступности.
:::tip
Мы протестировали выбранную группу моделей и провайдеров, которые хорошо работают с OpenCode.
:::
Чтобы исправить это, мы сделали пару вещей:
1. Мы протестировали выбранную группу открытых моделей и поговорили с их командами о том, как
лучше всего их запускать.
2. Затем мы поработали с несколькими провайдерами, чтобы убедиться, что они предоставляются
корректно.
3. Наконец, мы провели бенчмаркинг комбинаций модель/провайдер и составили
список, который мы смело можем рекомендовать.
OpenCode Go дает вам доступ к этим моделям за **$5 в первый месяц**, далее **$10 в месяц**.
---
## Как это работает
OpenCode Go работает так же, как и любой другой провайдер в OpenCode.
1. Вы входите в **<a href={console}>OpenCode Zen</a>**, подписываетесь на Go и
копируете свой API-ключ.
2. Вы выполняете команду `/connect` в TUI, выбираете `OpenCode Go` и вставляете
свой API-ключ.
3. Запустите `/models` в TUI, чтобы увидеть список моделей, доступных через Go.
:::note
Только один участник рабочего пространства может подписаться на OpenCode Go.
:::
Текущий список моделей включает:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Список моделей может меняться по мере того, как мы тестируем и добавляем новые.
---
## Лимиты использования
OpenCode Go включает следующие лимиты:
- **Лимит на 5 часов** — $12 использования
- **Недельный лимит** — $30 использования
- **Месячный лимит** — $60 использования
Лимиты определены в долларовом эквиваленте. Это означает, что ваше фактическое количество запросов зависит от используемой модели. Более дешевые модели, такие как DeepSeek V4 Flash, позволяют делать больше запросов, в то время как более дорогие, такие как GLM-5.2, — меньше.
В таблице ниже приведено примерное количество запросов на основе типичных сценариев использования Go:
| Model | запросов за 5 часов | запросов в неделю | запросов в месяц |
| ----------------- | ------------------- | ----------------- | ---------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Эти оценки основаны на наблюдаемых показателях запросов:
- Grok 4.5 — 1,100 входных, 71,500 кешированных, 220 выходных токенов на запрос
- GLM-5.2/5.1 — 700 входных, 52,000 кешированных, 150 выходных токенов на запрос
- Kimi K3 — 1,050 входных, 76,500 кешированных, 300 выходных токенов на запрос
- Kimi K2.7/K2.6 — 870 входных, 55,000 кешированных, 200 выходных токенов на запрос
- DeepSeek V4 Pro — 750 входных, 82,000 кешированных, 290 выходных токенов на запрос
- DeepSeek V4 Flash — 790 входных, 68,000 кешированных, 280 выходных токенов на запрос
- MiniMax M3 — 510 входных, 56,000 кешированных, 190 выходных токенов на запрос
- MiniMax M2.7 — 300 входных, 55,000 кешированных, 125 выходных токенов на запрос
- Qwen3.7 Max — 420 входных, 66,000 кешированных, 200 выходных токенов на запрос
- Qwen3.7 Plus — 500 входных, 57,000 кешированных, 190 выходных токенов на запрос
- Qwen3.6 Plus — 500 входных, 57,000 кешированных, 190 выходных токенов на запрос
- Hy3 — 830 входных, 71,500 кешированных, 295 выходных токенов на запрос
- MiMo-V2.5 — 830 входных, 71,500 кешированных, 295 выходных токенов на запрос
- MiMo-V2.5-Pro — 790 входных, 86,000 кешированных, 305 выходных токенов на запрос
Эти оценки также основаны на следующих ценах за 1M токенов и месячном объеме использования, включенном для каждой модели:
| Model | Input | Output | Cached Read | Cached Write | Использование |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ------------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Вы можете отслеживать текущее использование в **<a href={console}>консоли</a>**.
:::tip
Если вы достигнете лимита использования, вы можете продолжить использовать бесплатные модели.
:::
Лимиты использования могут измениться по мере того, как мы будем собирать данные о раннем использовании и отзывы.
---
### Использование сверх лимитов
Если у вас также есть средства на балансе Zen, вы можете включить опцию **Use balance**
в консоли. Если она включена, Go будет использовать ваш баланс Zen
после достижения лимитов использования вместо того, чтобы блокировать запросы.
---
### Почему для некоторых моделей доступен меньший объем использования
С Go вы платите $10 в месяц, а мы стремимся предоставить вам объем использования моделей стоимостью в шесть раз больше этой суммы.
Для большинства моделей это возможно благодаря оптовым скидкам и зарезервированным мощностям GPU. Полученную экономию мы передаем вам за счет шестикратного множителя.
Для некоторых моделей у нас еще не было возможности договориться о скидке или развернуть их с меньшими затратами: либо модель новая, либо ее публичные тарифы уже включают скидку.
Для этих моделей вы все равно получаете немного больше, чем при прямой оплате их провайдерам; именно поэтому указанный для них в таблице выше множитель использования ниже.
---
## Эндпоинты
Вы также можете получить доступ к моделям Go через следующие API-эндпоинты.
| Модель | ID модели | Эндпоинт | Пакет AI SDK |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
[ID модели](/docs/config/#models) в вашем конфиге OpenCode
использует формат `opencode-go/<model-id>`. Например, для Kimi K3 вам нужно
использовать `opencode-go/kimi-k3` в вашем конфиге.
---
### Модели
Вы можете получить полный список доступных моделей и их метаданных по адресу:
```
https://opencode.ai/zen/go/v1/models
```
---
## Конфиденциальность
Этот план разработан в первую очередь для международных пользователей: модели размещены в US, EU и Singapore, чтобы обеспечить стабильный доступ по всему миру. Наши провайдеры придерживаются политики zero-retention и не используют ваши данные для обучения моделей.
---
## Цели
Мы создали OpenCode Go, чтобы:
1. Сделать программирование с ИИ **доступным** для большего числа людей с помощью недорогой подписки.
2. Обеспечить **надежный** доступ к лучшим открытым моделям для программирования.
3. Отбирать модели, которые **протестированы и проверены** для использования в качестве агентов-программистов.
4. Избежать **привязки к провайдеру**, позволяя вам также использовать любого другого провайдера вместе с OpenCode.

224
packages/web/src/content/docs/th/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: การสมัครสมาชิกราคาประหยัดสำหรับโมเดลโอเพนซอร์สเพื่อการเขียนโค้ด
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go คือการสมัครสมาชิกในราคาประหยัด — **$5 สำหรับเดือนแรก** จากนั้น **$10/เดือน** — ซึ่งให้คุณเข้าถึงโมเดลโอเพนซอร์สยอดนิยมสำหรับการเขียนโค้ดได้อย่างเสถียร
Go ทำงานเหมือนกับผู้ให้บริการ (provider) รายอื่นๆ ใน OpenCode คุณสามารถสมัครสมาชิก OpenCode Go และรับ API key ของคุณ บริการนี้เป็น**ทางเลือกเพิ่มเติม** และคุณไม่จำเป็นต้องใช้มันเพื่อใช้งาน OpenCode
บริการนี้ออกแบบมาเพื่อผู้ใช้ในระดับสากลเป็นหลัก โดยมีโมเดลโฮสต์อยู่ในสหรัฐอเมริกา สหภาพยุโรป และสิงคโปร์ เพื่อการเข้าถึงทั่วโลกที่เสถียร
---
## Background
โมเดลแบบเปิด (Open models) ได้รับการพัฒนาจนมีประสิทธิภาพดีมาก ตอนนี้ประสิทธิภาพในการเขียนโค้ดใกล้เคียงกับโมเดลแบบปิด (proprietary models) แล้ว และเนื่องจากมีผู้ให้บริการหลายรายสามารถให้บริการแข่งขันกันได้ จึงทำให้มีราคาถูกกว่ามาก
อย่างไรก็ตาม การเข้าถึงโมเดลเหล่านี้อย่างเสถียรและมีค่า Latency ต่ำนั้นอาจเป็นเรื่องยาก ผู้ให้บริการแต่ละรายมีคุณภาพและความพร้อมในการให้บริการแตกต่างกันไป
:::tip
เราได้ทดสอบกลุ่มโมเดลและผู้ให้บริการที่คัดเลือกมาเพื่อให้ทำงานร่วมกับ OpenCode ได้เป็นอย่างดี
:::
เพื่อแก้ปัญหานี้ เราได้ดำเนินการดังต่อไปนี้:
1. เราทดสอบกลุ่มโมเดลแบบเปิดที่คัดเลือกมาและพูดคุยกับทีมผู้พัฒนาเกี่ยวกับวิธีการรันโมเดลเหล่านี้ให้ดีที่สุด
2. จากนั้นเราได้ทำงานร่วมกับผู้ให้บริการบางรายเพื่อให้แน่ใจว่าการให้บริการเป็นไปอย่างถูกต้อง
3. สุดท้าย เราได้ทำการวัดประสิทธิภาพ (benchmark) ของการทำงานร่วมกันระหว่างโมเดลและผู้ให้บริการ จนได้รายชื่อที่เรามั่นใจในการแนะนำ
OpenCode Go ให้คุณเข้าถึงโมเดลเหล่านี้ได้ในราคา **$5 สำหรับเดือนแรก** จากนั้น **$10/เดือน**
---
## How it works
OpenCode Go ทำงานเหมือนกับผู้ให้บริการรายอื่นๆ ใน OpenCode
1. ลงชื่อเข้าใช้ที่ **<a href={console}>OpenCode Zen</a>**, สมัครสมาชิก Go และคัดลอก API key ของคุณ
2. รันคำสั่ง `/connect` ใน TUI, เลือก `OpenCode Go`, และวาง API key ของคุณ
3. รัน `/models` ใน TUI เพื่อดูรายชื่อโมเดลที่ใช้งานได้ผ่าน Go
:::note
มีเพียงสมาชิกหนึ่งคนต่อ workspace เท่านั้นที่สามารถสมัครสมาชิก OpenCode Go ได้
:::
รายชื่อโมเดลในปัจจุบันประกอบด้วย:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
รายชื่อโมเดลอาจมีการเปลี่ยนแปลงเมื่อเราทำการทดสอบและเพิ่มโมเดลใหม่ๆ
---
## Usage limits
OpenCode Go มีขีดจำกัดดังต่อไปนี้:
- **ขีดจำกัดต่อ 5 ชั่วโมง** — การใช้งานมูลค่า $12
- **ขีดจำกัดรายสัปดาห์** — การใช้งานมูลค่า $30
- **ขีดจำกัดรายเดือน** — การใช้งานมูลค่า $60
ขีดจำกัดถูกกำหนดเป็นมูลค่าดอลลาร์ ซึ่งหมายความว่าจำนวน request จริงของคุณจะขึ้นอยู่กับโมเดลที่คุณใช้งาน โมเดลที่ราคาถูกกว่าอย่าง DeepSeek V4 Flash จะสามารถส่ง request ได้มากกว่า ในขณะที่โมเดลที่มีราคาสูงกว่าอย่าง GLM-5.2 จะส่งได้น้อยกว่า
ตารางด้านล่างแสดงจำนวน request โดยประมาณตามรูปแบบการใช้งานปกติของ Go:
| Model | requests ต่อ 5 ชั่วโมง | requests ต่อสัปดาห์ | requests ต่อเดือน |
| ----------------- | ---------------------- | ------------------- | ----------------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
การประมาณการนี้อ้างอิงจากรูปแบบการใช้งาน request ที่สังเกตพบ:
- Grok 4.5 — 1,100 input, 71,500 cached, 220 output tokens ต่อ request
- GLM-5.2/5.1 — 700 input, 52,000 cached, 150 output tokens ต่อ request
- Kimi K3 — 1,050 input, 76,500 cached, 300 output tokens ต่อ request
- Kimi K2.7/K2.6 — 870 input, 55,000 cached, 200 output tokens ต่อ request
- DeepSeek V4 Pro — 750 input, 82,000 cached, 290 output tokens ต่อ request
- DeepSeek V4 Flash — 790 input, 68,000 cached, 280 output tokens ต่อ request
- MiniMax M3 — 510 input, 56,000 cached, 190 output tokens ต่อ request
- MiniMax M2.7 — 300 input, 55,000 cached, 125 output tokens ต่อ request
- Qwen3.7 Max — 420 input, 66,000 cached, 200 output tokens ต่อ request
- Qwen3.7 Plus — 500 input, 57,000 cached, 190 output tokens ต่อ request
- Qwen3.6 Plus — 500 input, 57,000 cached, 190 output tokens ต่อ request
- Hy3 — 830 input, 71,500 cached, 295 output tokens ต่อ request
- MiMo-V2.5 — 830 input, 71,500 cached, 295 output tokens ต่อ request
- MiMo-V2.5-Pro — 790 input, 86,000 cached, 305 output tokens ต่อ request
การประมาณการนี้ยังอ้างอิงจากราคาต่อ 1M tokens และปริมาณการใช้งานรายเดือนที่รวมอยู่ในแต่ละโมเดลดังต่อไปนี้:
| Model | Input | Output | Cached Read | Cached Write | Usage |
| ---------------------------- | ------ | ------ | ----------- | ------------ | ----- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
คุณสามารถติดตามการใช้งานปัจจุบันของคุณได้ใน **<a href={console}>console</a>**
:::tip
หากคุณใช้งานถึงขีดจำกัดแล้ว คุณยังสามารถใช้งานโมเดลฟรีต่อไปได้
:::
ขีดจำกัดการใช้งานอาจมีการเปลี่ยนแปลงเมื่อเราได้เรียนรู้จากการใช้งานในช่วงแรกและรับฟังข้อเสนอแนะ
---
### Usage beyond limits
หากคุณมีเครดิตในยอดเงินคงเหลือของ Zen ด้วย คุณสามารถเปิดใช้งานตัวเลือก **Use balance** ใน console เมื่อเปิดใช้งาน Go จะหันไปใช้ยอดเงินคงเหลือใน Zen ของคุณหลังจากที่คุณใช้งานถึงขีดจำกัดแล้วแทนที่จะบล็อก request
---
### เหตุใดบางโมเดลจึงมีปริมาณการใช้งานต่ำกว่า
สำหรับ Go คุณจ่าย $10/เดือน และเราตั้งเป้าที่จะมอบปริมาณการใช้งานให้คุณ 6 เท่าของจำนวนดังกล่าว
สำหรับโมเดลส่วนใหญ่ เราทำเช่นนี้ได้ผ่านส่วนลดสำหรับการซื้อจำนวนมากและความจุ GPU ที่จองไว้ จากนั้นเราจะส่งต่อส่วนลดเหล่านั้นให้คุณผ่านตัวคูณ 6 เท่า
สำหรับบางโมเดล เรายังไม่มีโอกาสเจรจาส่วนลดหรือโฮสต์โมเดลเหล่านั้นด้วยต้นทุนที่ต่ำลง ไม่ว่าจะเป็นเพราะโมเดลนั้นใหม่หรือราคาที่ประกาศต่อสาธารณะมีส่วนลดอยู่แล้ว
สำหรับโมเดลเหล่านี้ คุณยังคงได้รับปริมาณการใช้งานมากกว่าการจ่ายให้ผู้ให้บริการโมเดลโดยตรงเล็กน้อย นี่คือเหตุผลที่ตัวคูณการใช้งานของโมเดลเหล่านี้ต่ำกว่าในตารางด้านบน
---
## Endpoints
คุณสามารถเข้าถึงโมเดลของ Go ผ่าน API endpoints ต่อไปนี้ได้เช่นกัน
| Model | Model ID | Endpoint | AI SDK Package |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
[model id](/docs/config/#models) ใน OpenCode config ของคุณจะใช้รูปแบบ `opencode-go/<model-id>` ตัวอย่างเช่น สำหรับ Kimi K3 คุณจะใช้ `opencode-go/kimi-k3` ใน config ของคุณ
---
### Models
คุณสามารถดึงรายการโมเดลทั้งหมดที่พร้อมใช้งานและ metadata ของมันได้จาก:
```
https://opencode.ai/zen/go/v1/models
```
---
## Privacy
แพลนนี้ออกแบบมาสำหรับผู้ใช้ทั่วโลกเป็นหลัก โดยโฮสต์โมเดลไว้ใน US, EU และ Singapore เพื่อให้เข้าถึงได้อย่างเสถียรจากทั่วโลก ผู้ให้บริการของเราปฏิบัติตามนโยบาย zero-retention และไม่นำข้อมูลของคุณไปใช้ในการฝึกโมเดล
---
## Goals
เราสร้าง OpenCode Go เพื่อ:
1. ทำให้การเขียนโค้ดด้วย AI **เข้าถึงได้** สำหรับผู้คนจำนวนมากขึ้นด้วยการสมัครสมาชิกในราคาประหยัด
2. มอบการเข้าถึงโมเดลโอเพนซอร์สสำหรับการเขียนโค้ดที่ดีที่สุดได้อย่าง**เสถียร**
3. คัดสรรโมเดลที่ผ่าน**การทดสอบและวัดประสิทธิภาพ**สำหรับการใช้งาน coding agent แล้ว
4. **ไม่ผูกขาดการใช้งาน (no lock-in)** โดยอนุญาตให้คุณใช้งานผู้ให้บริการรายอื่นๆ ร่วมกับ OpenCode ได้เช่นกัน

224
packages/web/src/content/docs/tr/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: Açık kodlama modelleri için düşük maliyetli abonelik.
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go, popüler açık kodlama modellerine güvenilir erişim sağlayan düşük maliyetli bir aboneliktir — **ilk ayınız için 5$**, sonrasında **aylık 10$**.
Go, OpenCode'daki diğer sağlayıcılar gibi çalışır. OpenCode Go'ya abone olur ve API anahtarınızı alırsınız. Bu **tamamen isteğe bağlıdır** ve OpenCode'u kullanmak için bunu kullanmanıza gerek yoktur.
Dünya çapında istikrarlı erişim için ABD, AB ve Singapur'da barındırılan modellerle temel olarak uluslararası kullanıcılar için tasarlanmıştır.
---
## Arka Plan
Açık modeller gerçekten çok iyi hale geldi. Artık kodlama görevleri için özel mülk modellere yakın bir performansa ulaşıyorlar. Ve birçok sağlayıcı onlara rekabetçi bir şekilde hizmet verebildiği için genellikle çok daha ucuzdurlar.
Ancak, bu modellere güvenilir, düşük gecikmeli erişim sağlamak zor olabilir. Sağlayıcılar kalite ve erişilebilirlik açısından farklılık gösterir.
:::tip
OpenCode ile iyi çalışan seçili bir model ve sağlayıcı grubunu test ettik.
:::
Bunu çözmek için birkaç şey yaptık:
1. Seçili bir grup açık modeli test ettik ve bu modelleri en iyi nasıl çalıştırabileceğimizi öğrenmek için ekipleriyle görüştük.
2. Ardından, bunların doğru şekilde sunulduğundan emin olmak için birkaç sağlayıcıyla birlikte çalıştık.
3. Son olarak, model/sağlayıcı kombinasyonunu kıyasladık (benchmark) ve gönül rahatlığıyla önerebileceğimiz bir liste oluşturduk.
OpenCode Go, bu modellere **ilk ayınız için 5$**, ardından **aylık 10$** karşılığında erişmenizi sağlar.
---
## Nasıl çalışır?
OpenCode Go, OpenCode'daki diğer sağlayıcılar gibi çalışır.
1. **<a href={console}>OpenCode Zen</a>**'e giriş yapın, Go'ya abone olun ve API anahtarınızı kopyalayın.
2. TUI'de `/connect` komutunu çalıştırın, `OpenCode Go`yu seçin ve API anahtarınızı yapıştırın.
3. Go üzerinden kullanabileceğiniz modellerin listesini görmek için TUI'de `/models` komutunu çalıştırın.
:::note
Her çalışma alanından yalnızca bir üye OpenCode Go'ya abone olabilir.
:::
Mevcut model listesi şunları içerir:
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
Test edip yenilerini ekledikçe model listesi değişebilir.
---
## Kullanım limitleri
OpenCode Go aşağıdaki limitleri içerir:
- **5 saatlik limit** — 12$ kullanım
- **Haftalık limit** — 30$ kullanım
- **Aylık limit** — 60$ kullanım
Limitler dolar değeri üzerinden belirlenmiştir. Bu, gerçek istek sayınızın kullandığınız modele bağlı olduğu anlamına gelir. DeepSeek V4 Flash gibi daha ucuz modeller daha fazla isteğe izin verirken, GLM-5.2 gibi yüksek maliyetli modeller daha azına izin verir.
Aşağıdaki tablo, tipik Go kullanım modellerine dayalı tahmini bir istek sayısı sunmaktadır:
| Model | 5 saatte bir istek | haftalık istek | aylık istek |
| ----------------- | ------------------ | -------------- | ----------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
Tahminler, gözlemlenen istek modellerine dayanır:
- Grok 4.5 — İstek başına 1.100 girdi, 71.500 önbelleğe alınmış, 220 çıktı token'ı
- GLM-5.2/5.1 — İstek başına 700 girdi, 52.000 önbelleğe alınmış, 150 çıktı token'ı
- Kimi K3 — İstek başına 1.050 girdi, 76.500 önbelleğe alınmış, 300 çıktı token'ı
- Kimi K2.7/K2.6 — İstek başına 870 girdi, 55.000 önbelleğe alınmış, 200 çıktı token'ı
- DeepSeek V4 Pro — İstek başına 750 girdi, 82.000 önbelleğe alınmış, 290 çıktı token'ı
- DeepSeek V4 Flash — İstek başına 790 girdi, 68.000 önbelleğe alınmış, 280 çıktı token'ı
- MiniMax M3 — İstek başına 510 girdi, 56.000 önbelleğe alınmış, 190 çıktı token'ı
- MiniMax M2.7 — İstek başına 300 girdi, 55.000 önbelleğe alınmış, 125 çıktı token'ı
- Qwen3.7 Max — İstek başına 420 girdi, 66.000 önbelleğe alınmış, 200 çıktı token'ı
- Qwen3.7 Plus — İstek başına 500 girdi, 57.000 önbelleğe alınmış, 190 çıktı token'ı
- Qwen3.6 Plus — İstek başına 500 girdi, 57.000 önbelleğe alınmış, 190 çıktı token'ı
- Hy3 — İstek başına 830 girdi, 71.500 önbelleğe alınmış, 295 çıktı token'ı
- MiMo-V2.5 — İstek başına 830 girdi, 71.500 önbelleğe alınmış, 295 çıktı token'ı
- MiMo-V2.5-Pro — İstek başına 790 girdi, 86.000 önbelleğe alınmış, 305 çıktı token'ı
Tahminler ayrıca 1M token başına aşağıdaki fiyatlara ve her modelle birlikte sunulan aylık kullanıma dayanır:
| Model | Input | Output | Cached Read | Cached Write | Kullanım |
| ---------------------------- | ------ | ------ | ----------- | ------------ | -------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
Mevcut kullanımınızı **<a href={console}>konsoldan</a>** takip edebilirsiniz.
:::tip
Kullanım sınırına ulaşırsanız, ücretsiz modelleri kullanmaya devam edebilirsiniz.
:::
Erken kullanımlardan ve geri bildirimlerden edindiğimiz deneyimlere bağlı olarak kullanım limitleri değişebilir.
---
### Limit dışı kullanım
Eğer Zen bakiyenizde kredileriniz varsa, konsoldan **Bakiye kullan (Use balance)** seçeneğini etkinleştirebilirsiniz. Bu özellik etkinleştirildiğinde, kullanım limitlerinize ulaştıktan sonra Go, istekleri engellemek yerine Zen bakiyenizi kullanmaya devam edecektir.
---
### Bazı modellerin kullanımı neden daha düşük?
Go ile aylık 10$ ödersiniz ve size bunun 6 katı değerinde kullanım sunmayı hedefleriz.
Çoğu modelde bunu toplu indirimler ve ayrılmış GPU kapasitesi sayesinde mümkün kılıyoruz. Ardından bu tasarrufları 6 katlık çarpanla size aktarıyoruz.
Bazı modellerde ise model yeni olduğu veya herkese açık fiyatlandırması zaten indirimli olduğu için henüz indirim pazarlığı yapma ya da modeli daha düşük maliyetle barındırma fırsatımız olmadı.
Bu modellerde bile model sağlayıcılarına doğrudan ödeme yaptığınız duruma kıyasla biraz daha fazla kullanım elde edersiniz; bu nedenle yukarıdaki tabloda kullanım çarpanları daha düşüktür.
---
## Uç Noktalar
Go modellerine aşağıdaki API uç noktaları aracılığıyla da erişebilirsiniz.
| Model | Model ID | Uç Nokta | AI SDK Paketi |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
OpenCode yapılandırmanızdaki [model id](/docs/config/#models) formatı `opencode-go/<model-id>` şeklindedir. Örneğin, Kimi K3 için yapılandırmanızda `opencode-go/kimi-k3` kullanmalısınız.
---
### Modeller
Mevcut modellerin tam listesini ve metadata'larını şuradan alabilirsiniz:
```
https://opencode.ai/zen/go/v1/models
```
---
## Gizlilik
Plan, öncelikle uluslararası kullanıcılar için tasarlanmıştır; dünya genelinde istikrarlı erişim sağlamak için modeller US, EU ve Singapore'da barındırılır. Sağlayıcılarımız sıfır veri saklama politikası uygular ve verilerinizi model eğitimi için kullanmaz.
---
## Hedefler
OpenCode Go'yu şu amaçlarla oluşturduk:
1. Düşük maliyetli bir abonelik ile yapay zeka destekli kodlamayı daha fazla kişi için **erişilebilir** kılmak.
2. En iyi açık kodlama modellerine **güvenilir** erişim sağlamak.
3. Kodlama aracısı (agent) kullanımı için **test edilmiş ve kıyaslanmış** modelleri bir araya getirmek.
4. OpenCode ile diğer sağlayıcıları da kullanmanıza olanak tanıyarak **kilitlenme (vendor lock-in) olmamasını** sağlamak.

224
packages/web/src/content/docs/zh-cn/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: 低成本的开源编程模型订阅服务。
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go 是一项低成本的订阅服务 —— **首月 5 美元**，之后 **每月 10 美元** —— 让你能够稳定地访问流行的开源编程模型。
Go 的工作方式与 OpenCode 中的任何其他提供商（provider）一样。订阅 OpenCode Go 后你将获得 API 密钥。它是 **完全可选** 的，并非使用 OpenCode 所必需的条件。
它主要为国际用户设计，模型托管在美国、欧盟和新加坡，以确保稳定的全球访问。
---
## 背景
开源模型现在变得非常强大。在编程任务中，它们的性能已接近专有模型。由于许多提供商都可以提供具有竞争力的服务，它们通常要便宜得多。
然而，获得可靠、低延迟的访问可能很困难。各提供商在质量和可用性方面参差不齐。
:::tip
我们测试了一组经过精选且与 OpenCode 配合良好的模型和提供商。
:::
为了解决这个问题，我们做了以下几件事：
1. 我们测试了一组精选的开源模型，并与他们的团队探讨了如何以最佳方式运行它们。
2. 随后我们与一些提供商合作，以确保正确提供这些服务。
3. 最后，我们对模型和提供商的组合进行了基准测试（benchmark），得出了一份我们乐于推荐的列表。
OpenCode Go 让你能够访问这些模型，**首月只需 5 美元**，之后 **每月 10 美元**。
---
## 工作原理
OpenCode Go 的工作方式与 OpenCode 中的其他提供商一样。
1. 登录 **<a href={console}>OpenCode Zen</a>**，订阅 Go，然后复制你的 API 密钥。
2. 在 TUI 中运行 `/connect` 命令，选择 `OpenCode Go`，然后粘贴你的 API 密钥。
3. 在 TUI 中运行 `/models` 以查看通过 Go 可用的模型列表。
:::note
每个工作空间只能有一名成员订阅 OpenCode Go。
:::
当前支持的模型列表包括：
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
随着我们进行测试和添加新模型，该列表可能会发生变化。
---
## 使用限制
OpenCode Go 包含以下限制：
- **5 小时限制** — 12 美元使用额度
- **每周限制** — 30 美元使用额度
- **每月限制** — 60 美元使用额度
限制以美元价值定义。这意味着你的实际请求数取决于你所使用的模型。较便宜的模型（如 DeepSeek V4 Flash）允许更多请求，而较高成本的模型（如 GLM-5.2）允许较少请求。
下表提供了基于典型 Go 使用模式的预估请求数：
| Model | 每 5 小时请求数 | 每周请求数 | 每月请求数 |
| ----------------- | --------------- | ---------- | ---------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
预估值基于观察到的请求模式：
- Grok 4.5 — 每次请求 1,100 个输入 token，71,500 个缓存 token，220 个输出 token
- GLM-5.2/5.1 — 每次请求 700 个输入 token，52,000 个缓存 token，150 个输出 token
- Kimi K3 — 每次请求 1,050 个输入 token，76,500 个缓存 token，300 个输出 token
- Kimi K2.7/K2.6 — 每次请求 870 个输入 token，55,000 个缓存 token，200 个输出 token
- DeepSeek V4 Pro — 每次请求 750 个输入 token，82,000 个缓存 token，290 个输出 token
- DeepSeek V4 Flash — 每次请求 790 个输入 token，68,000 个缓存 token，280 个输出 token
- MiMo-V2.5 — 每次请求 830 个输入 token，71,500 个缓存 token，295 个输出 token
- MiMo-V2.5-Pro — 每次请求 790 个输入 token，86,000 个缓存 token，305 个输出 token
- MiniMax M3 — 每次请求 510 个输入 token，56,000 个缓存 token，190 个输出 token
- MiniMax M2.7 — 每次请求 300 个输入 token，55,000 个缓存 token，125 个输出 token
- Qwen3.7 Max — 每次请求 420 个输入 token，66,000 个缓存 token，200 个输出 token
- Qwen3.7 Plus — 每次请求 500 个输入 token，57,000 个缓存 token，190 个输出 token
- Qwen3.6 Plus — 每次请求 500 个输入 token，57,000 个缓存 token，190 个输出 token
- Hy3 — 每次请求 830 个输入 token，71,500 个缓存 token，295 个输出 token
预估值还基于以下每 1M tokens 的价格以及每个模型包含的每月使用额度：
| 模型 | 输入 | 输出 | 缓存读取 | 缓存写入 | 使用额度 |
| ---------------------------- | ------ | ------ | --------- | -------- | -------- |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
你可以在 **<a href={console}>控制台</a>** 中跟踪你当前的使用情况。
:::tip
如果你达到了使用限制，你可以继续使用免费模型。
:::
使用限制可能会随着我们从早期使用和反馈中学习而发生变化。
---
### 超出限制的使用
如果你的 Zen 余额中还有积分，可以在控制台中启用 **使用余额（Use balance）** 选项。启用后，当你达到使用限制时，Go 会回退使用你的 Zen 余额，而不是拦截请求。
---
### 为什么某些模型的使用额度较低
使用 Go 时，你每月支付 $10，而我们的目标是为你提供 6 倍于此的使用额度。
对于大多数模型，我们通过批量折扣和预留 GPU 容量来实现这一目标。然后，我们通过 6 倍乘数将这些节省的成本回馈给你。
对于某些模型，我们还没有机会协商折扣或以更低的成本托管它们，这可能是因为模型较新，或者其公开价格已经是折扣价。
对于这些模型，你获得的使用额度仍会略高于直接向模型提供商付费；这就是它们在上表中的使用额度乘数较低的原因。
---
## API 端点
你也可以通过以下 API 端点访问 Go 模型。
| 模型 | 模型 ID | 端点 | AI SDK 包 |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
你的 OpenCode 配置中的 [模型 ID](/docs/config/#models) 使用 `opencode-go/<model-id>` 格式。例如，对于 Kimi K3，你将在配置中使用 `opencode-go/kimi-k3`。
---
### 模型
你可以从以下地址获取可用模型及其元数据的完整列表：
```
https://opencode.ai/zen/go/v1/models
```
---
## 隐私保护
该方案主要面向国际用户，模型托管在 US、EU 和 Singapore，以提供稳定的全球访问。我们的提供商遵循零保留政策，不会将您的数据用于模型训练。
---
## 目标
我们创建 OpenCode Go 的目的是：
1. 通过低成本订阅让更多人能够 **无门槛地** 使用 AI 编程。
2. 为最佳开源编程模型提供 **可靠的** 访问。
3. 精选经过 **测试和基准评估**，适合编程 Agent 使用的模型。
4. **无锁定（no lock-in）**，允许你与 OpenCode 一起使用任何其他提供商。

224
packages/web/src/content/docs/zh-tw/go.mdx
View file

@ -1,224 +0,0 @@
---
title: Go
description: 專為開源寫程式模型提供的低成本訂閱服務。
---
import config from "../../../../config.mjs"
export const console = config.console
export const email = `mailto:${config.email}`
OpenCode Go 是一項低成本的訂閱服務——**首月 $5 美元**，之後**每月 $10 美元**——讓您能穩定使用受歡迎的開源寫程式模型。
Go 的運作方式與 OpenCode 中的任何其他供應商相同。您訂閱 OpenCode Go 並取得您的 API key。這是**完全可選的**，您不需要使用它也能使用 OpenCode。
它主要為國際使用者設計，模型託管於美國、歐盟和新加坡，以提供全球穩定的存取。
---
## 背景
開源模型已經變得非常強大。它們在寫程式任務上的表現，現在已經接近專有模型。而且因為有許多供應商競爭提供服務，它們通常便宜得多。
然而，要獲得穩定、低延遲的存取可能很困難。各家供應商的品質和可用性不一。
:::tip
我們測試了一組精選的模型和供應商，它們與 OpenCode 搭配運作良好。
:::
為了解決這個問題，我們做了幾件事：
1. 我們測試了一組精選的開源模型，並與它們的團隊討論如何最佳化運行。
2. 接著我們與幾家供應商合作，確保這些模型被正確地提供服務。
3. 最後，我們對模型與供應商的組合進行了基準測試，並整理出一份我們樂於推薦的清單。
OpenCode Go 讓您可以存取這些模型，**首月只需 $5 美元**，之後**每月 $10 美元**。
---
## 運作方式
OpenCode Go 的運作方式與 OpenCode 中的任何其他供應商相同。
1. 您登入 **<a href={console}>OpenCode Zen</a>**，訂閱 Go，然後複製您的 API key。
2. 您在 TUI 中執行 `/connect` 命令，選擇 `OpenCode Go`，然後貼上您的 API key。
3. 在 TUI 中執行 `/models` 即可查看透過 Go 可用的模型清單。
:::note
每個工作區只能有一位成員訂閱 OpenCode Go。
:::
目前的模型清單包括：
- **Grok 4.5**
- **GLM-5.2**
- **GLM-5.1**
- **Kimi K3**
- **Kimi K2.7 Code**
- **Kimi K2.6**
- **MiMo-V2.5**
- **MiMo-V2.5-Pro**
- **MiniMax M3**
- **MiniMax M2.7**
- **Qwen3.7 Max**
- **Qwen3.7 Plus**
- **Qwen3.6 Plus**
- **DeepSeek V4 Pro**
- **DeepSeek V4 Flash**
- **Hy3**
隨著我們測試並加入新模型，模型清單可能會有所變動。
---
## 使用限制
OpenCode Go 包含以下限制：
- **5 小時限制** — $12 美元的使用量
- **每週限制** — $30 美元的使用量
- **每月限制** — $60 美元的使用量
限制是以美元價值來定義。這意味著您的實際請求次數取決於您使用的模型。像 DeepSeek V4 Flash 這樣較便宜的模型允許更多的請求次數，而像 GLM-5.2 這樣成本較高的模型則允許較少次數。
下表提供了基於典型 Go 使用模式的預估請求次數：
| Model | 每 5 小時請求數 | 每週請求數 | 每月請求數 |
| ----------------- | --------------- | ---------- | ---------- |
| Grok 4.5 | 120 | 300 | 600 |
| GLM-5.2 | 880 | 2,150 | 4,300 |
| GLM-5.1 | 880 | 2,150 | 4,300 |
| Kimi K3 | 110 | 250 | 490 |
| Kimi K2.7 Code | 1,350 | 3,380 | 6,750 |
| Kimi K2.6 | 1,150 | 2,880 | 5,750 |
| MiMo-V2.5 | 30,100 | 75,200 | 150,400 |
| MiMo-V2.5-Pro | 3,250 | 8,150 | 16,300 |
| MiniMax M3 | 3,200 | 8,000 | 16,000 |
| MiniMax M2.7 | 3,400 | 8,500 | 17,000 |
| Qwen3.7 Max | 950 | 2,390 | 4,770 |
| Qwen3.7 Plus | 4,300 | 10,800 | 21,600 |
| Qwen3.6 Plus | 3,300 | 8,200 | 16,300 |
| DeepSeek V4 Pro | 3,450 | 8,550 | 17,150 |
| DeepSeek V4 Flash | 31,650 | 79,050 | 158,150 |
| Hy3 | 4,300 | 10,750 | 21,500 |
這些預估值是基於觀察到的請求模式：
- Grok 4.5 — 每次請求 1,100 個輸入 token、71,500 個快取 token、220 個輸出 token
- GLM-5.2/5.1 — 每次請求 700 個輸入 token、52,000 個快取 token、150 個輸出 token
- Kimi K3 — 每次請求 1,050 個輸入 token、76,500 個快取 token、300 個輸出 token
- Kimi K2.7/K2.6 — 每次請求 870 個輸入 token、55,000 個快取 token、200 個輸出 token
- DeepSeek V4 Pro — 每次請求 750 個輸入 token、82,000 個快取 token、290 個輸出 token
- DeepSeek V4 Flash — 每次請求 790 個輸入 token、68,000 個快取 token、280 個輸出 token
- MiniMax M3 — 每次請求 510 個輸入 token、56,000 個快取 token、190 個輸出 token
- MiniMax M2.7 — 每次請求 300 個輸入 token、55,000 個快取 token、125 個輸出 token
- Qwen3.7 Max — 每次請求 420 個輸入 token、66,000 個快取 token、200 個輸出 token
- Qwen3.7 Plus — 每次請求 500 個輸入 token、57,000 個快取 token、190 個輸出 token
- Qwen3.6 Plus — 每次請求 500 個輸入 token、57,000 個快取 token、190 個輸出 token
- Hy3 — 每次請求 830 個輸入 token、71,500 個快取 token、295 個輸出 token
- MiMo-V2.5 — 每次請求 830 個輸入 token、71,500 個快取 token、295 個輸出 token
- MiMo-V2.5-Pro — 每次請求 790 個輸入 token、86,000 個快取 token、305 個輸出 token
這些預估值也基於以下每 1M tokens 的價格，以及每個模型所包含的每月使用量：
| 模型 | 輸入 | 輸出 | 快取讀取 | 快取寫入 | 使用量 |
| ---------------------------- | ------ | ------ | --------- | -------- | ------ |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | - | $15 |
| GLM-5.2 | $1.40 | $4.40 | $0.26 | - | $60 |
| GLM-5.1 | $1.40 | $4.40 | $0.26 | - | $60 |
| Kimi K3 | $3.00 | $15.00 | $0.30 | - | $15 |
| Kimi K2.7 Code | $0.95 | $4.00 | $0.19 | - | $60 |
| Kimi K2.6 | $0.95 | $4.00 | $0.16 | - | $60 |
| MiMo V2.5 | $0.14 | $0.28 | $0.0028 | - | $60 |
| MiMo V2.5 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | - | $60 |
| MiniMax M2.7 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| MiniMax M2.5 | $0.30 | $1.20 | $0.06 | $0.375 | $60 |
| Qwen3.7 Max | $2.50 | $7.50 | $0.50 | $3.125 | $60 |
| Qwen3.7 Plus (≤ 256K tokens) | $0.40 | $1.60 | $0.04 | $0.50 | $60 |
| Qwen3.7 Plus (> 256K tokens) | $1.20 | $4.80 | $0.12 | $1.50 | $60 |
| Qwen3.6 Plus (≤ 256K tokens) | $0.50 | $3.00 | $0.05 | $0.625 | $60 |
| Qwen3.6 Plus (> 256K tokens) | $2.00 | $6.00 | $0.20 | $2.50 | $60 |
| DeepSeek V4 Pro | $0.435 | $0.87 | $0.003625 | - | $15 |
| DeepSeek V4 Flash | $0.14 | $0.28 | $0.0028 | - | $60 |
| Hy3 | $0.14 | $0.58 | $0.035 | - | $60 |
您可以在 **<a href={console}>console</a>** 中追蹤您目前的使用量。
:::tip
如果您達到了使用限制，您可以繼續使用免費模型。
:::
使用限制可能會隨著我們從早期使用情況和回饋中學習而有所調整。
---
### 超出限制的使用量
如果您在您的 Zen 餘額中也有額度，您可以在 console 中啟用 **Use balance** 選項。啟用後，當您達到使用限制時，Go 將會改用您的 Zen 餘額，而不是阻擋請求。
---
### 為什麼部分模型的使用量較低
使用 Go 時，您每月支付 $10，而我們的目標是提供相當於 6 倍費用的使用量。
對大多數模型而言，我們透過大量採購折扣和預留 GPU 容量來達成此目標，再以 6 倍乘數將節省的成本回饋給您。
對於部分模型，我們尚未有機會議定折扣或以較低成本託管，原因可能是模型剛推出，或其公開價格已經過折扣。
對於這些模型，您獲得的使用量仍會比直接向模型供應商付費多一些；這就是上表中其使用量乘數較低的原因。
---
## 端點
您也可以透過以下 API 端點存取 Go 模型。
| 模型 | 模型 ID | 端點 | AI SDK 套件 |
| ----------------- | ----------------- | ------------------------------------------------ | --------------------------- |
| Grok 4.5 | grok-4.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.2 | glm-5.2 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| GLM-5.1 | glm-5.1 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K3 | kimi-k3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.7 Code | kimi-k2.7-code | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| Kimi K2.6 | kimi-k2.6 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Pro | deepseek-v4-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| DeepSeek V4 Flash | deepseek-v4-flash | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5 | mimo-v2.5 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiMo-V2.5-Pro | mimo-v2.5-pro | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
| MiniMax M3 | minimax-m3 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.7 | minimax-m2.7 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| MiniMax M2.5 | minimax-m2.5 | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Max | qwen3.7-max | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.7 Plus | qwen3.7-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Qwen3.6 Plus | qwen3.6-plus | `https://opencode.ai/zen/go/v1/messages` | `@ai-sdk/anthropic` |
| Hy3 | hy3 | `https://opencode.ai/zen/go/v1/chat/completions` | `@ai-sdk/openai-compatible` |
您的 OpenCode 設定中的 [model id](/docs/config/#models) 使用 `opencode-go/<model-id>` 格式。例如，Kimi K3 在設定中應使用 `opencode-go/kimi-k3`。
---
### 模型
你可以從以下位置取得所有可用模型及其中繼資料的完整清單：
```
https://opencode.ai/zen/go/v1/models
```
---
## 隱私權
此方案主要為國際使用者設計，模型部署於 US、EU 與 Singapore，以提供穩定的全球存取體驗。我們的供應商遵循零保留政策，不會將你的資料用於模型訓練。
---
## 目標
我們建立 OpenCode Go 的目的是：
1. 透過低成本的訂閱服務，讓更多人能**輕鬆獲得** AI 寫程式的協助。
2. 提供**穩定可靠**的途徑來存取最好的開源寫程式模型。
3. 精選經過**測試與效能評估**、適合寫程式代理使用的模型。
4. 允許您在 OpenCode 中同時使用任何其他供應商，確保**不被單一廠商綁定**。

2
packages/web/src/middleware.ts
View file

@ -78,6 +78,8 @@ function localeFromAcceptLanguage(header: string | null) {
}
export const onRequest = defineMiddleware((ctx, next) => {
if (/^\/docs\/[^/]+\/go\/?$/.test(ctx.url.pathname)) return redirect(ctx.url, "/docs/go")
const alias = docsAlias(ctx.url.pathname)
if (alias) {
return redirect(ctx.url, alias.path, alias.locale)

15
packages/www/content/docs/(Configure)/providers.mdx
View file

@ -57,6 +57,21 @@ When several credential sources exist, OpenCode uses the stored credential first
<Callout type="warning">Do not commit API keys or authorization headers to your repository.</Callout>
## OpenCode Console
OpenCode Console uses browser device authorization for a named user account:
```bash
opencode console login
```
After authorization, OpenCode loads the Console-managed model configuration into the `opencode` provider. An explicit
Console server URL may be passed for a custom deployment; the default is `https://opencode.ai/console`.
[OpenCode Go](/go) subscriptions use this named-user flow and the `opencode` provider. Do not copy an API key. API-key
and service-account connections remain available for ordinary managed inference, but they are not eligible for Go
subscription access.
## Configure
The `providers` object is keyed by provider ID. Each provider accepts these fields:

36
packages/www/content/docs/go.mdx Normal file
View file

@ -0,0 +1,36 @@
---
title: "OpenCode Go"
description: "Connect a named OpenCode Go subscription to OpenCode."
---
OpenCode Go is a low-cost subscription for popular open coding models. Go access belongs to the named person who
subscribes and is available through OpenCode's `opencode` provider.
<Callout type="note">
OpenCode Go does not use a copied API key. Service accounts and other coding agents are not eligible.
</Callout>
## Subscribe
Subscribe in [OpenCode Console](https://opencode.ai/console/go). The subscription is attached to your Console account.
## Connect OpenCode
Run the Console device login:
```bash
opencode console login
```
OpenCode opens a browser authorization page. Approve the device for your Console account. The CLI then loads the
`opencode` provider configuration for your organization.
For the initial launch, an account with exactly one organization connects automatically. If your account has no
organization or belongs to multiple organizations, login stops with instructions instead of choosing one implicitly.
## Use Go models
Choose a model from the `opencode` provider in OpenCode. No separate Go provider or endpoint configuration is required.
API keys and service-account credentials can still connect ordinary OpenCode managed inference. They do not grant
OpenCode Go subscription access.

6
packages/www/content/docs/index.mdx
View file

@ -104,10 +104,10 @@ Run `/connect` in the TUI and select your provider.
```
If you'd like easy access to all the best coding models you can try out
[OpenCode Console](https://console.opencode.ai).
[OpenCode Console](https://opencode.ai/console).
You can also try [OpenCode Go](https://opencode.ai/go) a $10/month subscription
plan that grants you access to the best open source models.
You can also try [OpenCode Go](/go), a subscription for a named OpenCode user that grants access to popular open coding
models through the `opencode` provider.
Use `/models` to browse the providers and models available to your project. See [Providers](/providers) for connection and
configuration details.