MXFP8 Open on Hugging Face
Parameters
123B
Weights (est.)
≈ 131 GB
Mac memory (est.)
192 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-vlm on an Apple silicon Mac:
pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/Qwen3.5-122B-A10B-mxfp8 --image photo.jpg --prompt "Describe this image"
Details
- Base model
- Qwen/Qwen3.5-122B-A10B (quantized)
- Vendor
- Qwen
- Task
- Vision + text
- Quantisation
- MXFP8
- Licence
- apache-2.0
- Downloads
- 791 in the last 30 days · 2.7K all time · 0 likes
- Published
- 2026-05-02 · updated 2026-05-02 (6mo ago)
- Tags
- qwen3_5_moe
Other MLX builds of Qwen/Qwen3.5-122B-A10B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen3.5-122B-A10B-OptiQ-2bit | 2-bit | ≈ 39.1 GB | 48 GB+ | 353 |
| Qwen3.5-122B-A10B-heretic-v2-2.34bit-msq | 34-bit | ≈ 43.3 GB | 64 GB+ | 84 |
| Qwen3.5-122B-A10B-mxfp4 | MXFP4 | ≈ 69.6 GB | 96 GB+ | 126 |
| Qwen3.5-122B-A10B-6bit-gs32 | 6-bit | ≈ 100 GB | 128 GB+ | 86 |
| Qwen3.5-122B-A10B-mxfp8 (this one) | MXFP8 | ≈ 131 GB | 192 GB+ | 791 |