Gemma 4 12B: A unified, encoder-free multimodal model

· HN · Frontpage AI ·

Google released Gemma 4 12B as a unified multimodal model, signaling continued progress in smaller open-weight vision-language systems.

Categories: Model Releases

Excerpt

HN · 267 points · 98 comments

Read at source: https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemma-4-12b/

Discussions