Google Launches Gemini 3.1 Ultra With 2M Token Context and Native Multimodal

Google's Gemini 3.1 Ultra ships with a 2-million token context window operating natively across text, image, audio, and video — no transcription intermediaries required. A sandboxed Code Execution tool is built in, letting the model write and run code mid-conversation. The model is the headline infrastructure release of May 2026.

Source: AI Tools Recap