Google has introduced Gemini Omni, a new AI model designed to process and understand multiple types of data simultaneously, including text, images, audio, and video. The model aims to enable more natural and intuitive interactions by combining these modalities in a single framework, building on Google's ongoing work in multimodal AI research.
At Google I/O, Google announced Gemini Spark, a personal AI agent running on Gemini 3.5 Flash and Antigravity. The open source Gemini CLI will stop working with AI subscriptions on June 18th, replaced by a closed source Antigravity CLI.
Google released Gemini 3.5 Flash directly to general availability, rolling it out across its Gemini app, Search, and developer platforms. The model is 3x the price of 3 Flash Preview, fitting a broader trend of AI labs raising API prices. A new beta Interactions API with server-side history management was also announced.
Google announced Gemini Spark, a new AI personal agent capable of navigating users' digital lives and acting on their behalf across Google products. The agent has been tested with limited users and will be available next week to subscribers of AI Ultra, a new $100-per-month tier.
llm-gemini 0.32a0 has been released. It is compatible with the llm>=0.32a0 alpha and adds the ability to stream reasoning tokens.
The release of llm-gemini 0.32 adds support for the new gemini-3.5-flash model, enabling users to access Gemini 3.5 Flash through the plugin.