Google’s Gemini 3 Breaks Records With New Coding App
▼ Summary
– Google released Gemini 3, its most advanced foundation model, which is now available through the Gemini app and AI search interface.
– The model demonstrates a major leap in reasoning capabilities, achieving the highest score on record on the Humanity’s Last Exam benchmark.
– Gemini 3 has over 650 million monthly active users in the app and is used by 13 million software developers in their workflows.
– Google also launched Antigravity, a Gemini-powered coding interface that combines a prompt window, command-line, and browser for agentic coding.
– A research-intensive version called Gemini 3 Deepthink will be available to Google AI Ultra subscribers after further safety testing.
Google has officially launched Gemini 3, its most powerful foundation model to date, now accessible through the Gemini app and AI search interface. This release arrives just seven months after Gemini 2.5, positioning it as a formidable competitor in the rapidly advancing AI market. The timing is notable, coming shortly after OpenAI’s GPT 5.1 and Anthropic’s Sonnet 4.5, underscoring the intense speed of development in cutting-edge AI models.
A specialized research-focused iteration, Gemini 3 Deepthink, will be rolled out to Google AI Ultra subscribers in the coming weeks, pending the completion of additional safety evaluations.
Tulsee Doshi, Google’s head of product for Gemini, emphasized the model’s significant advancements. She noted that Gemini 3 demonstrates a massive leap in reasoning capabilities, delivering responses with an unprecedented level of depth and subtlety.
Independent benchmarks are already reflecting this enhanced performance. The model achieved a record-breaking score of 37.4 on the Humanity’s Last Exam, a benchmark designed to assess general reasoning and expertise. This surpasses the previous high of 31.64 held by GPT-5 Pro. Additionally, Gemini 3 claimed the top position on LMArena, a human-led benchmark that gauges user satisfaction.
Google reports that the Gemini app now boasts over 650 million monthly active users, with 13 million software developers incorporating the model into their daily workflows.
Accompanying the core model, Google introduced Google Antigravity, a new coding interface powered by Gemini. This tool enables multi-pane agentic coding, similar to platforms like Warp or Cursor 2.0. Antigravity uniquely integrates a ChatGPT-style prompt window with a command-line interface and a browser window, allowing developers to see the real-time impact of changes made by the coding agent.
DeepMind CTO Koray Kavukcuoglu explained that the agent is designed to collaborate across a developer’s editor, terminal, and browser to ensure applications are built as effectively as possible.
(Source: TechCrunch)
