---

base_model:

• Qwen/Qwen3.8-Flash-Next

base_model_relation: quantized

license: other

license_name: qwen-community-1.0

license_link: LICENSE

library_name: gguf

pipeline_tag: text-generation

tags:

• gguf
• mtp
• speculative-decoding
• qwen4exp

---

Qwen3.8-Flash-Next MTP draft head (Q8_0)

The multi-token-prediction head from

Qwen/Qwen3.8-Flash-Next, 3.85 GiB at Q8_0.

Qwen trains it jointly with the model, so it drafts better than a separate small model

would. Use it as the draft for any Qwen3.8-Flash-Next GGUF.

Setup

Qwen3.8-Flash-Next itself has been in upstream llama.cpp since b10660, but MTP for it has

not, so this head still needs the fork. On upstream you get:

error loading model: check_tensor_dims: tensor 'blk.0.hc_attn_norm.weight' not found

Upstream's qwen4exp loader has no MTP-only path, so it requires the 48 trunk layers this

file does not carry. Build this branch instead:

git clone https://github.com/LaurentZuijdwijk/llama.cpp
cd llama.cpp && git checkout vulkan/qwen4exp-rocmfpx
cmake -B build -DGGML_VULKAN=ON -DCMAKE_BUILD_TYPE=Release
cmake --build build -j

Run

./build/bin/llama-server \
  -m Qwen3.8-Flash-Next.gguf \
  -md Qwen3.8-Flash-Next-MTP-Q8_0.gguf \
  -ngl 99 --n-gpu-layers-draft 99 \
  --spec-type draft-mtp --spec-draft-n-max 3 -c 32768

Adds ~3.9 GiB on top of the target.

Measurements

Radeon 8060S, 250 tokens at temp 0, warmed up first.

| target | no draft | with this head | acceptance |

|---|---|---|---|

| UD-IQ4_XS | 24.2 t/s | 29.3 t/s | 0.623 |

| ROCmFP4-FAST | 28.1 t/s | 30.3 t/s | 0.587 |

A 4-bit version of this head

measured faster on both targets and is 1.6 GiB smaller.

Credits

Qwen3.8-Flash-Next support in llama.cpp is the work of Daniel Han

(@danielhanchen), from

ggml-org/llama.cpp#27742; the MTP graph

is from #27739 (JJJYmmm). Base model by

the Qwen team.

Quantized and published by Agention.

License

Qwen Community License 1.0, included as LICENSE.