Google releases Gemini 3 AI model
The major update to Google’s flagship model features better reasoning, multistep agentic tasks, and stronger “vibe coding.”
Today Google released Gemini 3, a major update to its flagship AI model. The new model comes with a number of new features that aim to leap ahead of competitors OpenAI and Anthropic. The release also features significant improvements to reasoning, agentic tasks, and “vibe coding.” Google says when using Gemini 3, “you’ll notice the responses are more helpful, better formatted and more concise.”
Google says that Gemini’s interface has also been refreshed, and the shopping experience has “radically” improved.
Gemini 3 has a new feature called Generative Interfaces, which can turn responses into magazine-style layouts or build interactive presentations as a response to queries, though both of these examples were listed as “experiments.”
Another experiment is Gemini Agent, which can handle complex, multistep tasks across Google’s apps all within Gemini.
Google hopes to leverage its massive reach so Gemini becomes the AI chatbot of choice. The company posted results that show Gemini bested GPT-5.1 and Claude Sonnet 4.5 in many widely used benchmarks. Gemini 3 leapt to the top spot on the popular LMArena leaderboard.
