feat(moe): Qwen3.8-Flash-Next support (#172)

Qwen3.8-Flash-Next (qwen4exp): 125B total, ~6B active, 512 routed experts at
top-10 plus one shared, 48 hybrid gated-delta SSM / sparse attention layers,
and a 51B n-gram embedding table. One registry row streams the experts; a
dense-policy guard keeps the n-gram table (larger than any phone's RAM)
mmap'd under every mode so pinned and anonymous dense weights survive load.
Runs on the 12 GB test phone at ~2 tok/s with pinned dense weights, compute-
bound, and sits in the app catalog as a three-shard download. Submodule
pinned to upstream master b10666, the first with the architecture merged,
with the expert-ready hook on top. README hero clip, changelog and docs
updated. App 0.22.0 (versionCode 37).

examples/android/README.md

@@ -95,7 +95,7 @@ check). Nothing below needs a storage permission except the last option.
    adb shell mv /data/local/tmp/shardllm /data/local/tmp/bmoe
    ```
 
-### Sharded models (gpt-oss-120b, DeepSeek V4 Flash)
+### Sharded models (gpt-oss-120b, DeepSeek V4 Flash, Qwen3.8-Flash-Next)
 
 Models above Hugging Face's 50 GB per-file limit ship as several shard files
 (`-00001-of-0000N.gguf`). The engine streams a split set natively, so these download in-app
@@ -111,7 +111,13 @@ the first one:
 adb push DeepSeek-V4-Flash-0731-UD-IQ2_M-0000*-of-00003.gguf /data/local/tmp/bmoe/
 ```
 
-Mind the space: DeepSeek V4 Flash UD-IQ2_M is ~91 GB on disk.
+Mind the space: DeepSeek V4 Flash UD-IQ2_M is ~91 GB on disk, Qwen3.8-Flash-Next UD-IQ3_XXS ~82 GB.
+
+Qwen3.8-Flash-Next needs **Dense weights = Pinned (dma-buf)** in Settings. Its dense side is 4.3 GB
+and every token walks it: with Anon the kernel swaps it to zram and single tokens stall for 10-20 s,
+with Mmap it is refaulted from flash every token. Its 51B n-gram table is held back automatically and
+stays mmap'd whatever the setting; the engine says so on stderr at load. Keep the expert cache at
+1000-1500 MiB on a 12 GB phone: the pinned dense set leaves no room for more.
 
 ## Expected numbers
 

examples/android/app/build.gradle

@@ -51,8 +51,8 @@ android {
         applicationId 'io.bigmoeonedge.example'
         minSdk 29
         targetSdk 34
-        versionCode 36
-        versionName '0.21.0'
+        versionCode 37
+        versionName '0.22.0'
         buildConfigField 'String', 'GIT_SHA', "\"${gitSha}\""
         ndk {
             // The engine ships as prebuilt arm64 binaries staged by build-android.ps1.

examples/android/app/src/main/java/io/bigmoeonedge/example/ModelCatalog.kt

@@ -153,6 +153,37 @@ object ModelCatalog {
                 ),
             ),
         ),
+        Entry(
+            title = "Qwen3.8-Flash-Next",
+            quant = "UD-IQ3_XXS",
+            fileName = "Qwen3.8-Flash-Next-UD-IQ3_XXS-00001-of-00003.gguf",
+            approxBytes = 81_961_823_936L,
+            url = null,
+            // The dense side is 4.3 GB and is walked every token, so on a 12 GB phone this one
+            // runs only with the dense weights pinned (dma-buf); anon swaps, mmap refaults. The
+            // 51B n-gram table stays mmap'd on its own — see docs/android-memory.md.
+            blurb = "6B active of 125B, the Qwen4 preview. ~82 GB on disk; needs Pinned dense weights.",
+            shards = listOf(
+                Shard(
+                    "Qwen3.8-Flash-Next-UD-IQ3_XXS-00001-of-00003.gguf",
+                    "https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/resolve/main/UD-IQ3_XXS/" +
+                        "Qwen3.8-Flash-Next-UD-IQ3_XXS-00001-of-00003.gguf?download=true",
+                    10_946_624L,
+                ),
+                Shard(
+                    "Qwen3.8-Flash-Next-UD-IQ3_XXS-00002-of-00003.gguf",
+                    "https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/resolve/main/UD-IQ3_XXS/" +
+                        "Qwen3.8-Flash-Next-UD-IQ3_XXS-00002-of-00003.gguf?download=true",
+                    49_567_921_344L,
+                ),
+                Shard(
+                    "Qwen3.8-Flash-Next-UD-IQ3_XXS-00003-of-00003.gguf",
+                    "https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF/resolve/main/UD-IQ3_XXS/" +
+                        "Qwen3.8-Flash-Next-UD-IQ3_XXS-00003-of-00003.gguf?download=true",
+                    32_382_955_968L,
+                ),
+            ),
+        ),
     )
 
     /**