GoogleGemma 3
Gemma 3 12B
A mid-size Gemma 3 vision-language checkpoint.
What the release provides
The 12B model accepts images and text, generates text, and has a shared 128K window. Instruction weights are published.Google Gemma 3 model card
How to use it well
Practical guidance. Compare this size with the smaller checkpoint using the same image preparation and output contract. Check whether added capacity reduces corrections on your difficult documents. Include the full context length and simultaneous request count in hardware measurements.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | google/gemma-3-12b-it |