MXFP4 Open on Hugging Face
Parameters
4.0B
Weights (est.)
≈ 2.3 GB
Mac memory (est.)
8 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/JOSIE-1.1-4B-Thinking-mxfp4 --prompt "Hello"
Details
- Base model
- Goekdeniz-Guelmez/JOSIE-1.1-4B-Thinking (quantized)
- Vendor
- Goekdeniz-Guelmez
- Task
- Text generation
- Quantisation
- MXFP4
- Licence
- mit
- Downloads
- 17 in the last 30 days · 122 all time · 0 likes
- Published
- 2026-03-09 · updated 2026-03-09 (8mo ago)
- Languages
- multilingual, en, es, fr, pt, it, ar, ko, id, ru, vi, de, th, ja, zh
- Tags
- qwen3, chat
Other MLX builds of Goekdeniz-Guelmez/JOSIE-1.1-4B-Thinking
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| JOSIE-1.1-4B-Thinking-mxfp4 (this one) | MXFP4 | ≈ 2.3 GB | 8 GB+ | 17 |
| JOSIE-1.1-4B-Thinking-4bit | 4-bit | ≈ 2.3 GB | 8 GB+ | 17 |
| JOSIE-1.1-4B-Thinking-6bit | 6-bit | ≈ 3.3 GB | 8 GB+ | 23 |
| JOSIE-1.1-4B-Thinking-8bit | 8-bit | ≈ 4.3 GB | 8 GB+ | 16 |
| JOSIE-1.1-4B-Thinking-bfloat16 | — | ≈ 8.0 GB | 16 GB+ | 23 |