Qwen3.6-35B-A3B
· llama.cpp · 128K context · Arc B390 · 64 GB
Qwen3.6's 35B mixture-of-experts model with 3B active parameters.
28 tok/sdecode
332 tok/sprefill
128Kcontext window
Tested on this cardOct 10, 2026
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 Download the weights
hf download unsloth/Qwen3.6-35B-A3B-GGUF Qwen3.6-35B-A3B-UD-Q4_K_M.gguf \ --revision a483e9e6cbd595906af30beda3187c2663a1118c \ --local-dir ~/models/Qwen3.6-35B-A3B-GGUF-a483e9e6
- 2 Start the server
docker run --rm \ --device /dev/dri \ -p 8000:8080 \ --shm-size 8g \ -v ~/models/Qwen3.6-35B-A3B-GGUF-a483e9e6:/models:ro \ ghcr.io/ggml-org/llama.cpp@sha256:86b37d266ce85b485127adea68c9dc6d04d0ec9a64b0979844d5bbd8d66c53c5 \ -m /models/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf \ --alias Qwen3.6-35B-A3B-GGUF \ --host 0.0.0.0 \ --port 8080 \ --ctx-size 131072 \ --n-gpu-layers 999 \ --parallel 1 \ --flash-attn on \ --jinja \ --reasoning-format deepseek \ --no-mmproj \ --metrics
✓ 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 B390 iGPU (Dell XPS 14, Core Ultra X7 358H, 64 GB shared; on battery, power-saver) on Oct 10, 2026.
- Weights
- unsloth/Qwen3.6-35B-A3B-GGUF @ a483e9e6cb ›
- Image
- ghcr.io/ggml-org/llama.cpp@sha256:86b37d266ce85b485127adea68c9dc6d04d0ec9a64b0979844d5bbd8d66c53c5
- Launch
- llama.cpp-sycl-qwen3.6-35b-a3b-ud-q4-k-m-128k ›
- Recipe file
- registry/recipes/intel/intel-arc-b390-64gb/qwen3.6-35b-a3b.llama.cpp.128k.json ›
- Model released
- Apr 15, 2026
- Tested
- owner, Intel Arc B390 iGPU (Dell XPS 14, Core Ultra X7 358H, 64 GB shared; on battery, power-saver)