Qwen3-4B-Thinking-2507-fp16
mlx-community/Qwen3-4B-Thinking-2507-fp16
Text generation · transformers
fp16 Open on Hugging Face
Parameters
4.0B
Weights (est.)
≈ 8.0 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/Qwen3-4B-Thinking-2507-fp16 --prompt "Hello"
Details
- Base model
- Qwen/Qwen3-4B-Thinking-2507 (finetune)
- Vendor
- Qwen
- Task
- Text generation
- Quantisation
- fp16
- Licence
- apache-2.0
- Downloads
- 38 in the last 30 days · 540 all time · 1 likes
- Published
- 2025-10-10 · updated 2025-10-10 (1y ago)
- Tags
- qwen3
Other MLX builds of Qwen/Qwen3-4B-Thinking-2507
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen3-4B-Thinking-2507-4bit | 4-bit | ≈ 2.3 GB | 8 GB+ | 891 |
| Qwen3-4B-Thinking-2507-5bit | 5-bit | ≈ 2.8 GB | 8 GB+ | 38 |
| Qwen3-4B-Thinking-2507-6bit | 6-bit | ≈ 3.3 GB | 8 GB+ | 31 |
| Qwen3-4B-Thinking-2507-8bit | 8-bit | ≈ 4.3 GB | 8 GB+ | 138 |
| Qwen3-4B-Thinking-2507-fp16 (this one) | fp16 | ≈ 8.0 GB | 16 GB+ | 38 |