Downloads · 30 days
10
9% of all-time downloads
Gemma-4-31B-it-EXL3-6.0bpw
HaoweiShen/Gemma-4-31B-it-EXL3-6.0bpw
Gemma-4-31B-it-EXL3-6.0bpw is a machine learning model from HaoweiShen. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for exllamav3. The card lists the license as gemma.
Quantized version of google/gemma-4-31B-it using ExLlamaV3 EXL3 format.
All-time downloads
115
Public
Parameters
13.5B
27 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors27 GB · 100%
Parameter types
How the weights are stored.
I1611.5B · 85%
Base models
Model card
From the Hugging Face model README
Gemma-4-31B-it EXL3 6.0bpw
Quantized version of google/gemma-4-31B-it using ExLlamaV3 EXL3 format.
| Property | Value |
|---|---|
| Original size | 62 GB (BF16) |
| Quantized size | 25 GB (6.0 bpw) |
| Format | EXL3 (QTIP-based) |
| Compression | 2.5x |
Requirements
- ExLlamaV3 with Gemma 4 support: SimonShenhw/exllamav3-gemma4
- GPU VRAM: 28+ GB
Credits
- Architecture adaptation: @lesj0610
- Inference fix + quantization: @SimonShenhw
- ExLlamaV3: @turboderp
- Library
- exllamav3
- Type
- gemma4
- License
- gemma
- Created
- Apr 7, 2026
- Updated
- Apr 7, 2026