Sybil AI
‹ Arc Pro B65

Gemma 4 31B

· llama.cpp · 128K context · Arc Pro B65 · 32 GB

Google's dense Gemma 4, 31B parameters.

21 tok/sdecode
326 tok/sprefill
128Kcontext window
Tested on this cardOct 14, 2026
Run it

Follow the steps below to download and run this recipe. Omarchy Local AI lists the recipes it can launch on your hardware. The server then answers on http://localhost:8000/v1.

  1. 1 Download the weights
    hf download lmstudio-community/gemma-4-31B-it-QAT-GGUF gemma-4-31B-it-QAT-Q4_0.gguf \
      --revision 12e5eae7a583b0a8bac91919fa99ae1fb27e6de4 \
      --local-dir ~/models/gemma-4-31B-it-QAT-GGUF-12e5eae7
  2. 2 Start the server
    docker run --rm \
      --device /dev/dri \
      -p 8000:8080 \
      --shm-size 8g \
      -e ONEAPI_DEVICE_SELECTOR=level_zero:0 \
      -e UR_L0_ENABLE_RELAXED_ALLOCATION_LIMITS=1 \
      -v ~/models/gemma-4-31B-it-QAT-GGUF-12e5eae7:/models:ro \
      ghcr.io/ggml-org/llama.cpp@sha256:86b37d266ce85b485127adea68c9dc6d04d0ec9a64b0979844d5bbd8d66c53c5 \
      -m /models/gemma-4-31B-it-QAT-Q4_0.gguf \
      --alias gemma-4-31B-it-QAT-GGUF \
      --host 0.0.0.0 \
      --port 8080 \
      --ctx-size 131072 \
      --n-gpu-layers 999 \
      --parallel 1 \
      --flash-attn on \
      --batch-size 2048 \
      --ubatch-size 2048 \
      --jinja \
      --reasoning-format deepseek \
      --no-mmproj \
      --metrics
What it passed
✓ loadLoads and serves within an hour
✓ chatAnswers a plain question and stops on its own
✓ reasoningThinks separately and gets 17 × 23 right
✓ toolsCalls a tool with the right arguments and uses the result
✓ contextRecalls a code buried in a prompt that fills 85% of the window
✓ speedDecodes at 15 tok/s or more

Passed all six checks on a real Intel Arc Pro B65 32 GB (ASRock Creator; owner: Ryzen 7 5700X, PCIe 4.0 x4) on Oct 14, 2026.

Details
Weights
lmstudio-community/gemma-4-31B-it-QAT-GGUF @ 12e5eae7a5 ›
Image
ghcr.io/ggml-org/llama.cpp@sha256:86b37d266ce85b485127adea68c9dc6d04d0ec9a64b0979844d5bbd8d66c53c5
Launch
llama.cpp-sycl-gemma-4-31b-qat-q4-0-128k-arc-pro-b65 ›
Recipe file
registry/recipes/intel/intel-arc-pro-b65-32gb/gemma-4-31b.llama.cpp.128k.json ›
Model released
Mar 11, 2026
Tested
owner, Intel Arc Pro B65 32 GB (ASRock Creator; owner: Ryzen 7 5700X, PCIe 4.0 x4)