leonsarmiento/Qwen3.8-27B-3bit-mlx
Image-Text-to-Text • 27B • Updated • 888 • 1
Qwen3.8-27B MLX Quantizations
3-bit VLM base + 4-bit MTP drafter. Pair both for accelerated instruct decode. reasoning_effort baked to low.
3-bit VLM base + 4-bit MTP drafter. Pair both for accelerated instruct decode. reasoning_effort baked to low.