4-bit Open on Hugging Face
Parameters
25B
Weights (est.)
≈ 15.3 GB
Mac memory (est.)
24 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/gemma-4-26B-A4B-it-OptiQ-4bit --prompt "Hello"
Details
- Base model
- google/gemma-4-26B-A4B-it (quantized)
- Vendor
- Google · Gemma
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- gemma
- Downloads
- 1.2K in the last 30 days · 32K all time · 21 likes
- Published
- 2026-04-25 · updated 2026-10-18 (24d ago)
- Tags
- gemma4, quantized, mixed-precision, 4bit, 8bit, optiq, apple-silicon, gemma-4
Other MLX builds of google/gemma-4-26B-A4B-it
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| gemma-4-26b-a4b-it-4bit | 4-bit | ≈ 15.3 GB | 24 GB+ | 17K |
| gemma-4-26b-a4b-it-nvfp4 | 4-bit | ≈ 15.3 GB | 24 GB+ | 371 |
| gemma-4-26b-a4b-it-mxfp4 | MXFP4 | ≈ 15.3 GB | 24 GB+ | 203 |
| gemma-4-26B-A4B-it-OptiQ-4bit (this one) | 4-bit | ≈ 15.3 GB | 24 GB+ | 1.2K |
| gemma-4-26b-a4b-it-5bit | 5-bit | ≈ 18.5 GB | 24 GB+ | 509 |
| gemma-4-26b-a4b-it-6bit | 6-bit | ≈ 21.6 GB | 32 GB+ | 541 |
| gemma-4-26B-A4B-it-heretic-msq-2.6bit | 6-bit | ≈ 21.7 GB | 32 GB+ | 392 |
| gemma-4-26b-a4b-it-8bit | 8-bit | ≈ 28.0 GB | 48 GB+ | 3.7K |
| gemma-4-26b-a4b-it-mxfp8 | MXFP8 | ≈ 28.0 GB | 48 GB+ | 135 |
| gemma-4-26b-a4b-it-bf16 | bf16 | ≈ 51.6 GB | 64 GB+ | 430 |