Alibaba Qwen · First MLX build 18 Oct 2026
Qwen2.5-Omni-3B
MLX
qwen-research
A model from Alibaba Qwen, with 3 MLX builds, from 4-bit to bf16.
- Parameters
- — Not published
- Quantisation
- 4-bit Qwen2.5-Omni-3B-Thinker-4bit
- Memory needed
- — GB Estimate, unified memory
- Download
- — Weights, estimated
- Downloads
- 492 This build, last 30 days
- Task
- —
- MLX builds
- 3 On mlx-community
- Devices
- Mac Apple silicon
Memory
Hugging Face does not list a weights size for this build, so there is no memory estimate.
Estimated from the size of the weights, plus a fifth for context and working memory and 2 GB for macOS and your apps. Long contexts need more.
Run it
Terminal · Hugging Face CLI$ pip install -U huggingface_hub
$ hf download mlx-community/Qwen2.5-Omni-3B-Thinker-4bitMLX builds
3 on mlx-community| Qwen2.5-Omni-3B-Thinker-4bit | 4-bit | — | — | 492 |
|---|---|---|---|---|
| Qwen2.5-Omni-3B-Thinker-8bit | 8-bit | — | — | 437 |
| Qwen2.5-Omni-3B-Thinker-bf16 | bf16 | — | — | 283 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.