MXFP8 Open on Hugging Face
Parameters
30B
Weights (est.)
≈ 31.8 GB
Mac memory (est.)
48 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/GLM-4.7-Flash-8bit-mxfp8 --prompt "Hello"
Details
- Base model
- zai-org/GLM-4.7-Flash (quantized)
- Vendor
- zai-org
- Task
- Text generation
- Quantisation
- MXFP8
- Licence
- mit
- Downloads
- 75 in the last 30 days · 2.2K all time · 1 likes
- Published
- 2026-02-10 · updated 2026-02-10 (9mo ago)
- Languages
- en, zh
- Tags
- glm4_moe_lite
Other MLX builds of zai-org/GLM-4.7-Flash
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| GLM-4.7-Flash-4bit | 4-bit | ≈ 16.9 GB | 24 GB+ | 3.8K |
| GLM-4.7-Flash-4bit-mxfp4 | MXFP4 | ≈ 16.9 GB | 24 GB+ | 181 |
| GLM-4.7-Flash-5bit | 5-bit | ≈ 20.6 GB | 32 GB+ | 745 |
| GLM-4.7-Flash-6bit | 6-bit | ≈ 24.3 GB | 32 GB+ | 991 |
| GLM-4.7-Flash-8bit | 8-bit | ≈ 31.8 GB | 48 GB+ | 1.3K |
| GLM-4.7-Flash-8bit-mxfp8 (this one) | MXFP8 | ≈ 31.8 GB | 48 GB+ | 75 |
| GLM-4.7-Flash-bf16 | bf16 | ≈ 59.9 GB | 96 GB+ | 93 |