4-bit Open on Hugging Face
Parameters
26B
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-vlm on an Apple silicon Mac:
pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/diffusiongemma-26B-A4B-it-nvfp4 --image photo.jpg --prompt "Describe this image"
Details
- Base model
- google/diffusiongemma-26B-A4B-it (quantized) · diffusiongemma 26B A4B it on the radar →
- Vendor
- Google · Gemma
- Task
- Vision + text
- Quantisation
- 4-bit
- Licence
- apache-2.0
- Downloads
- 192 in the last 30 days · 4.6K all time · 6 likes
- Published
- 2026-06-10 · updated 2026-07-15 (3mo ago)
- Tags
- diffusion_gemma
Other MLX builds of google/diffusiongemma-26B-A4B-it
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| diffusiongemma-26B-A4B-it-OptiQ-4bit | 4-bit | ≈ 14.2 GB | 24 GB+ | 534 |
| diffusiongemma-26B-A4B-it-4bit | 4-bit | ≈ 15.3 GB | 24 GB+ | 1.7K |
| diffusiongemma-26B-A4B-it-nvfp4 (this one) | 4-bit | ≈ 15.3 GB | 24 GB+ | 192 |
| diffusiongemma-26B-A4B-it-mxfp4 | MXFP4 | ≈ 15.3 GB | 24 GB+ | 152 |
| diffusiongemma-26B-A4B-it-5bit | 5-bit | ≈ 18.5 GB | 24 GB+ | 168 |
| diffusiongemma-26B-A4B-it-6bit | 6-bit | ≈ 21.7 GB | 32 GB+ | 141 |
| diffusiongemma-26B-A4B-it-8bit | 8-bit | ≈ 28.0 GB | 48 GB+ | 276 |
| diffusiongemma-26B-A4B-it-mxfp8 | MXFP8 | ≈ 28.0 GB | 48 GB+ | 77 |
| diffusiongemma-26B-A4B-it-bf16 | bf16 | ≈ 51.6 GB | 64 GB+ | 91 |