Alibaba Qwen · First MLX build 28 Apr 2025 · In Core AI since 2 Sept 2026
Qwen3-1.7B
MLXCore AI
Text apache-2.0
A 2B parameter text generation model from Alibaba Qwen, with 10 MLX builds, from 3-bit to bf16 and Apple Core AI export recipes for macOS and iOS.
- Parameters
- 2B From Hugging Face
- Quantisation
- 4-bit Qwen3-1.7B-4bit
- Memory needed
- 4 GB Estimate, unified memory
- Download
- 968 MB Weights, estimated
- Downloads
- 16K This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 10 On mlx-community
- Devices
- Mac Apple silicon
- Parameters
- 2B From Hugging Face
- Compression
- 4-bit, None Apple’s preset
- Memory needed
- 4 GB Estimate from parameters
- Context
- 32K tokens Fixed at export
- Precision
- FP16 Compute precision
- Platforms
- macOS, iOS Mac and iPhone
- Export presets
- 2 In Apple’s registry
- Listed by Apple
- 2 Sept 2026 First registry commit
Memory
Will it fit?Needs about 4 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/Qwen3-1.7B-4bit --prompt "Hello"Memory
Will it fit?Needs about 4 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 qwen3-1.7b
$ uv run coreai.llm.export qwen3-1.7b --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
10 on mlx-community| Qwen3-1.7B-3bit | 3-bit | 753 MB | 3 GB | 87 |
|---|---|---|---|---|
| Qwen3-1.7B-4bit | 4-bit | 968 MB | 4 GB | 16K |
| Qwen3-1.7B-4bit-DWQ-053125 | 4-bit DWQ | 968 MB | 4 GB | 842 |
| Qwen3-1.7B-4bit-DWQ | 4-bit DWQ | 968 MB | 4 GB | 293 |
| Qwen3-1.7B-MLX-MXFP4 | MXFP4 | 968 MB | 4 GB | 82 |
| Qwen3-1.7B-4bit-AWQ | 4-bit | 968 MB | 4 GB | 79 |
| Qwen3-1.7B-5bit | 5-bit | 1.2 GB | 4 GB | 72 |
| Qwen3-1.7B-6bit | 6-bit | 1.4 GB | 4 GB | 107 |
| Qwen3-1.7B-8bit | 8-bit | 1.8 GB | 5 GB | 769 |
| Qwen3-1.7B-bf16 | bf16 | 3.4 GB | 6 GB | 676 |
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 |
|---|---|---|---|---|
| qwen3-1.7b | macOS | 4-bit | FP16 | 32K |
| qwen3-1.7b | iOS | None | FP16 | 4K |