Hajra Naz
Google rolled out Gemini 3 on Tuesday. This is its newest and most advanced foundation model. It is live now in the Gemini app and the AI search interface.
The launch arrives only seven months after Gemini 2.5. Gemini 3 stands as Google’s strongest LLM so far. It enters the market as a direct rival to every leading generative AI system. The timing is sharp. OpenAI released GPT 5.1 less than a week ago. Anthropic pushed Sonnet 4.5 only two months earlier. AI updates continue at extreme speed across the frontier model race.
Google will also ship a more research-heavy version soon. The variant is named Gemini 3 Deepthink. It will reach Google AI Ultra subscribers once safety checks are complete.
Read More: Google Launches New Gemini Certifications to Boost AI Skills in Education
Tulsee Doshi, Google’s head of product for Gemini, pointed to a major rise in reasoning quality. Doshi said Gemini 3 now responds with depth and nuance not seen in past versions.
Independent evaluations already show that jump. Gemini 3 scored 37.4 on the Humanity’s Last Exam benchmark. It is now the highest score on record. GPT-5 Pro held the earlier top score at 31.64. Gemini 3 also leads on LMArena, a human-run benchmark focused on user satisfaction.
Google says the Gemini app now has more than 650 million monthly active users. Thirteen million developers use the model in their workflow. This expands the model’s reach inside software teams and across AI-driven coding pipelines.
Google also introduced a new coding interface. The tool is named Google Antigravity. It uses Gemini to power an agentic coding setup. The interface supports multipane work similar to Warp or Cursor 2.0. It blends a prompt window, a command-line panel, and a browser preview. This lets the coding agent apply updates and show results in real time.
Read More: Google Rolls Out Gemini AI on Google TV Streamer
DeepMind CTO Koray Kavukcuoglu said the agent connects across the editor, terminal, and browser. It supports the full build process for apps and coding projects. This places Antigravity into the broader wave of AI-assisted development tools now trending across the industry.
FAQs
1. What makes Gemini 3 different from Gemini 2.5?
Gemini 3 delivers stronger reasoning, higher benchmark scores, new agentic abilities, and wider integration through the Gemini app.
2. What is Gemini 3 Deepthink?
It is a research-focused version with deeper reasoning features. It will roll out to Google AI Ultra subscribers after more safety reviews.
3. How did Gemini 3 perform on AI benchmarks?
It reached a record 37.4 on Humanity’s Last Exam and topped user-rated results on LMArena.
4. What is Google Antigravity?
It is a Gemini-powered coding tool with multi-pane support for prompts, terminal work, and live browser previews.
5. Who can access Gemini 3 today?
It is available through the Gemini app and Google’s AI search interface. Deepthink will follow for Ultra subscribers.
