MXFP8 Open on Hugging Face
Parameters
423M
Weights (est.)
≈ 449 MB
Mac memory (est.)
8 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.
Details
- Base model
- google/gemma-4-12B-it-assistant (quantized)
- Vendor
- Google · Gemma
- Task
- Any to any
- Quantisation
- MXFP8
- Licence
- apache-2.0
- Downloads
- 37 in the last 30 days · 981 all time · 1 likes
- Published
- 2026-06-03 · updated 2026-06-03 (5mo ago)
- Tags
- gemma4_unified_assistant
Other MLX builds of google/gemma-4-12B-it-assistant
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| gemma-4-12B-it-assistant-4bit | 4-bit | ≈ 238 MB | 8 GB+ | 206 |
| gemma-4-12B-it-assistant-nvfp4 | 4-bit | ≈ 238 MB | 8 GB+ | 52 |
| gemma-4-12B-it-assistant-mxfp4 | MXFP4 | ≈ 238 MB | 8 GB+ | 44 |
| gemma-4-12B-it-assistant-5bit | 5-bit | ≈ 291 MB | 8 GB+ | 34 |
| gemma-4-12B-it-assistant-6bit | 6-bit | ≈ 344 MB | 8 GB+ | 42 |
| gemma-4-12B-it-assistant-8bit | 8-bit | ≈ 449 MB | 8 GB+ | 73 |
| gemma-4-12B-it-assistant-mxfp8 (this one) | MXFP8 | ≈ 449 MB | 8 GB+ | 37 |
| gemma-4-12B-it-assistant-bf16 | bf16 | ≈ 846 MB | 8 GB+ | 711 |