Views
No views yet
Qwen/Qwen3.8-27B using
mlx-lm.
Refer to the original model card
for more details on the model.
Weights: ~15 GB (4.501 effective bits/weight) on disk (bf16 MLX artifact: ~54 GB; the original
checkpoint is ~55 GB including the vision encoder and MTP drafter, which MLX
conversions drop).qwen3_5 implementation strips the
base model's vision encoder and multi-token-prediction drafter; this artifact
does not accept images or videos. For vision use, run the original checkpoint
with Transformers, vLLM, or SGLang.reasoning_effort control (xhigh default / medium / low) and
preserve_thinking. See the
base model card for details and
benchmarks.1pip install -U mlx-lm
2mlx_lm.generate --model malekoo/Qwen3.8-27B-MLX-4bit --prompt "Explain KV caches briefly." --max-tokens 2048mlx_lm.server --model malekoo/Qwen3.8-27B-MLX-4bit --port 8080.
Thinking control via chat_template_kwargs: enable_thinking,
preserve_thinking, reasoning_effort (xhigh/medium/low).temperature=1.0, top_p=0.95, top_k=20, min_p=0.0,
presence_penalty=0.0, repetition_penalty=1.0temperature=0.7, top_p=0.80, top_k=20,
min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0Qwen/Qwen3.8-27B (bf16 safetensors, released 2026-08-14)mlx_lm.convert -q --q-bits 4 --q-group-size 640.32.1.dev20260814+3d23f7d87, mlx-lm at ml-explore/mlx-lm main 254d1531@misc{qwen38,
2 title = {{Qwen3.8-Max}: A New Bar for Coding and Cowork},
3 url = {https://qwen.ai/blog?id=qwen3.8},
4 author = {{Qwen Team}},
5 month = {August},
6 year = {2026}
7}