Google unveils the next generation of its most powerful AI technologies
Google has introduced major updates in its AI ecosystem, announcing the Gemini 3 Pro model and the new AI-agent development platform Antigravity. The updates are now available to Google AI subscribers and some free-tier users.
Gemini 3 Pro: significantly more powerful
According to Google, Gemini 3 Pro outperforms Gemini 2.5 Pro in key benchmarks and provides more accurate analysis of user queries. The model is designed to deliver correct, essential information rather than simply expected answers.
In Google Search’s AI Mode, the model is activated through a dedicated menu. The company plans to make Gemini 3 Pro widely available to all U.S. users soon.
Improvements in the Gemini app
In the Gemini app, enabling the “Thinking” mode unlocks the advanced features of Gemini 3 Pro. Answer quality and Canvas application functionality have been noticeably improved. Users with Plus, Pro and Ultra plans will receive expanded usage limits.
Deep Think — advanced reasoning mode
Google also introduced Gemini 3 Deep Think, which scored highly in tests such as the Last Exam, GPQA Diamond, and ARC-AGI-2. Currently limited to safety testers, this feature will soon roll out to Google AI Ultra subscribers.
Antigravity — new AI-agent development platform
Google also launched Antigravity, a free public preview platform that allows AI agents to perform tasks and conduct code checks on behalf of users. The system is designed to automate and speed up development workflows.
