Google & DeepMind

Google unveils Gemini 4 Neon with a sharper focus on agentic work
Google introduces Gemini 4 Neon across developer and selected consumer products, highlighting a new generation of useful AI systems.

Gemini Decisions API aims at fast multimodal choices
Google introduced a model and endpoint for fast decisions in tasks where an app needs a bounded answer rather than an open-ended essay.

AlphaProof Code points to a new path for formally verified software
DeepMind says its researchers used an agentic verification framework to prove security properties in code connected with the Signal protocol.

Gemini Robotics 2.5 brings AI models closer to physical tasks
New robotics work focuses on general-purpose dexterity and the difficult gap between virtual reasoning and real-world action.

Genie 4 explores interactive worlds generated by AI
Google DeepMind introduced Genie 4 as a research preview focused on world simulation, with restricted early access.

Gemini 3.5 expands Google’s vision for coding agents
Google’s I/O 2026 keynote introduced a stronger model family designed for building and using agentic applications.