Gemma 4 12B brings multimodal AI into one smaller, encoder-free model
Google is rolling out Gemma 4 12B, positioning it as a unified, encoder-free multimodal model designed to simplify how developers build AI systems that handle more than just text.
Google shows off 9 demos of Gemini Omni and Gemini 3.5 in action
Google has released nine demo videos highlighting what Gemini Omni and Gemini 3.5 can do across multimodal understanding, real-time interaction, and practical AI workflows.
Gemini API File Search goes multimodal, pushing RAG beyond plain text
Google has expanded Gemini API File Search to multimodal retrieval, giving developers a way to build RAG systems that can pull grounded context from more than just text.