Google launches Gemini 3, its most intelligent AI model

Google has introduced Gemini 3, its most advanced AI model to date, featuring state-of-the-art reasoning and multimodal capabilities. The model tops the LMArena Leaderboard with a score of 1501 Elo and achieves PhD-level reasoning on Humanity's Last Exam (37.5% without tools) and GPQA Diamond (91.9%). Gemini 3 Pro is available immediately across Google products including the Gemini app, AI Studio, Vertex AI, and AI Mode in Search. A new Deep Think mode will be released to Google AI Ultra subscribers in the coming weeks after additional safety evaluations. Google also unveiled Google Antigravity, a new agentic development platform that leverages Gemini 3's advanced reasoning and tool use. The model demonstrates significant improvements in coding benchmarks, scoring 76.2% on SWE-bench Verified and 54.2% on Terminal-Bench 2.0. Gemini 3 has undergone the most comprehensive safety evaluations of any Google AI model, with partnerships including the UK AISI and independent assessments from Apollo, Vaultis, and Dreadnode. The announcement was made by Google and Alphabet CEO Sundar Pichai, DeepMind CEO Demis Hassabis, and DeepMind CTO Koray Kavukcuoglu.