Google launched Gemini 3.1 Pro on February 19, 2026, rolling it out globally through the Gemini app and NotebookLM in early April. The model scored 77.1% on ARC-AGI-2, up from 3 Pro's 31.1%, and reached 80.6% on SWE-Bench Verified, nearly matching Anthropic's Claude Opus 4.6. It supports a 1M-token context window with full multimodal input and is available to Gemini Pro and Ultra subscribers, via the Gemini API, and through Vertex AI for enterprise users.