Alibaba Qwen · First MLX build 28 Apr 2025 · In Core AI since 8 Jun 2026
Qwen3-8B
MLXCore AI
Text apache-2.0 Trending
A 8.2B parameter text generation model from Alibaba Qwen, with 9 MLX builds, from 3-bit to bf16 and Apple Core AI export recipes for macOS and iOS.
- Parameters
- 8.2B From Hugging Face
- Quantisation
- 4-bit Qwen3-8B-4bit
- Memory needed
- 8 GB Estimate, unified memory
- Download
- 4.6 GB Weights, estimated
- Downloads
- 33K This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 9 On mlx-community
- Devices
- Mac Apple silicon
- Parameters
- 8.2B From Hugging Face
- Compression
- 4-bit, None Apple’s preset
- Memory needed
- 8 GB Estimate from parameters
- Context
- 40K tokens Fixed at export
- Precision
- FP16 Compute precision
- Platforms
- macOS, iOS Mac and iPhone
- Export presets
- 2 In Apple’s registry
- Listed by Apple
- 8 Jun 2026 First registry commit
Memory
Will it fit?Needs about 8 GB at 4-bit
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
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 · mlx-lm$ pip install -U mlx-lm
$ mlx_lm.generate --model mlx-community/Qwen3-8B-4bit --prompt "Hello"Memory
Will it fit?Needs about 8 GB at 4-bit, None
8 GB
16 GB
24 GB
32 GB
64 GB
128 GB
Fits with headroomTight, close other appsWon’t fit
Estimated from the parameter count at Apple’s compression setting. Apple does not publish package sizes.
Export it
Terminal · uv$ git clone https://github.com/apple/coreai-models
$ cd coreai-models
$ uv run coreai.llm.export qwen3-8b
$ uv run coreai.llm.export qwen3-8b --platform iOS
Apple publishes the recipe, not a ready-made package: these commands convert the original weights to a Core AI .aimodel on your Mac. See the repository for requirements.
MLX builds
9 on mlx-community| Qwen3-8B-DFly-MLX | Unstated | 2.3 GB | 5 GB | 13 |
|---|---|---|---|---|
| Qwen3-8B-3bit | 3-bit | 3.6 GB | 7 GB | 193 |
| Qwen3-8B-4bit | 4-bit | 4.6 GB | 8 GB | 33K |
| Qwen3-8B-4bit-DWQ | 4-bit DWQ | 4.6 GB | 8 GB | 364 |
| Qwen3-8B-4bit-DWQ-053125 | 4-bit DWQ | 4.6 GB | 8 GB | 221 |
| Qwen3-8B-4bit-AWQ | 4-bit | 4.6 GB | 8 GB | 198 |
| Qwen3-8B-6bit | 6-bit | 6.7 GB | 10 GB | 345 |
| Qwen3-8B-8bit | 8-bit | 8.7 GB | 12 GB | 943 |
| Qwen3-8B-bf16 | bf16 | 16.4 GB | 21 GB | 577 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.
Core AI export presets
From Apple’s registry| Preset | Platform | Compression | Precision | Context |
|---|---|---|---|---|
| qwen3-8b | macOS | 4-bit | FP16 | 40K |
| qwen3-8b | iOS | None | FP16 | 4K |