Back to AI intel
重点

Google releases Gemma-4 12B, a unified multimodal model with tool calling and 256K context.

AI intel briefing

Core summary

One sentence to understand this update

Google has released Gemma-4 12B, described as a "beast" of a unified multimodal model with direct audio/image input, tool calling, and a 256K context window.

Impact & opportunity

What this could mean

Builders can leverage Gemma-4 12B's multimodal capabilities and large context for advanced applications, especially those requiring complex interactions and tool use.