Tencent · First MLX build 8 Feb 2025
Hunyuan-7B-Instruct
MLX
Text tencent-license
A 7.5B parameter text generation model from Tencent, with 7 MLX builds, from 3-bit to fp16.
- Parameters
- 7.5B From Hugging Face
- Quantisation
- 4-bit Hunyuan-7B-Instruct-2508-4bit
- Memory needed
- 7 GB Estimate, unified memory
- Download
- 4.2 GB Weights, estimated
- Downloads
- 45 This build, last 30 days
- Task
- Text generation Text
- MLX builds
- 7 On mlx-community
- Devices
- Mac Apple silicon
Memory
Will it fit?Needs about 7 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/Hunyuan-7B-Instruct-2508-4bit --prompt "Hello"MLX builds
7 on mlx-community| Hunyuan-7B-Instruct-3bit | 3-bit | 3.3 GB | 6 GB | 18 |
|---|---|---|---|---|
| Hunyuan-7B-Instruct-2508-4bit | 4-bit | 4.2 GB | 7 GB | 45 |
| Hunyuan-7B-Instruct-4bit | 4-bit | 4.2 GB | 7 GB | 41 |
| Hunyuan-7B-Instruct-6bit | 6-bit | 6.1 GB | 9 GB | 20 |
| Hunyuan-7B-Instruct-2508-8bit | 8-bit | 8.0 GB | 11 GB | 33 |
| Hunyuan-7B-Instruct-8bit | 8-bit | 8.0 GB | 11 GB | 24 |
| Hunyuan-7B-Instruct-fp16 | fp16 | 15.0 GB | 19 GB | 24 |
Sizes are estimated from Hugging Face’s parameter counts and the quantisation. Downloads are for the last 30 days.