Tencent · First MLX build 9 Jul 2026
Hy3
MLX
Text apache-2.0
A 295B parameter text generation model from Tencent, with 6 MLX builds at 2-bit.
- Parameters
- 295B From Hugging Face
- Quantisation
- 2-bit Hy3-OptiQ-2bit
- Memory needed
- 106 GB Estimate, unified memory
- Download
- 92.3 GB Weights, estimated
- Downloads
- 141 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 6 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 106 GB at 2-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/Hy3-OptiQ-2bit --prompt "Hello"MLX builds
6 on mlx-community| Hy3-OptiQ-2bit | 2-bit | 92.3 GB | 106 GB | 141 |
|---|---|---|---|---|
| Hy3-oQ2 | 2-bit | 92.3 GB | 106 GB | 82 |
| Hy3-oQ2e-2.31bpw | 2-bit | 92.3 GB | 106 GB | 80 |
| Hy3-oQ2e | 2-bit | 92.3 GB | 106 GB | 56 |
| Hy3-oQ2e-2.33bpw | 2-bit | 92.3 GB | 106 GB | 43 |
| Hy3-oQ2e-2.37bpw | 2-bit | 92.3 GB | 106 GB | 39 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.