MXFP4 Open on Hugging Face
Parameters
21B
Weights (est.)
≈ 11.8 GB
Mac memory (est.)
16 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-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/gpt-oss-20b-MXFP4-Q8 --prompt "Hello"
Details
- Base model
- openai/gpt-oss-20b (quantized) · gpt oss 20b on the radar →
- Vendor
- OpenAI · gpt-oss
- Task
- Text generation
- Quantisation
- MXFP4
- Licence
- apache-2.0
- Downloads
- 173K in the last 30 days · 7.9M all time · 101 likes
- Published
- 2025-08-29 · updated 2026-03-19 (7mo ago)
- Tags
- gpt_oss, vllm
Other MLX builds of openai/gpt-oss-20b
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| gpt-oss-20b-MXFP4-Q8 (this one) | MXFP4 | ≈ 11.8 GB | 16 GB+ | 173K |
| gpt-oss-20b-MXFP4-Q4 | MXFP4 | ≈ 11.8 GB | 16 GB+ | 2.3K |
| gpt-oss-20b-OptiQ-4bit | 4-bit | ≈ 11.8 GB | 16 GB+ | 1.8K |
| gpt-oss-20b-mxfp4-bf16 | MXFP4 | ≈ 14.4 GB | 24 GB+ | 442 |