6-bit Open on Hugging Face
Parameters
15B
Weights (est.)
≈ 11.9 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/Phi-4-reasoning-6bit --prompt "Hello"
Details
- Base model
- microsoft/Phi-4-reasoning (quantized) · Phi 4 reasoning on the radar →
- Vendor
- Microsoft · Phi
- Task
- Text generation
- Quantisation
- 6-bit
- Licence
- mit
- Downloads
- 32 in the last 30 days · 332 all time · 1 likes
- Published
- 2025-05-01 · updated 2025-05-01 (1y ago)
- Languages
- en
- Tags
- phi3, math, code, chat, reasoning
Other MLX builds of microsoft/Phi-4-reasoning
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Phi-4-reasoning-4bit | 4-bit | ≈ 8.2 GB | 16 GB+ | 68 |
| Phi-4-reasoning-6bit (this one) | 6-bit | ≈ 11.9 GB | 16 GB+ | 32 |
| Phi-4-reasoning-8bit | 8-bit | ≈ 15.6 GB | 24 GB+ | 40 |
| Phi-4-reasoning-bf16 | bf16 | ≈ 29.3 GB | 48 GB+ | 29 |