Google has announced the release of Gemini 3, its latest foundation model, through the Gemini app and AI search interface. This update follows the recent launch of Gemini 2.5 and demonstrates Google’s commitment to advancing AI capabilities.
Gemini 3 is Google’s most advanced Large Language Model (LLM) to date, positioning the company as a significant player in the competitive AI landscape. Its release, closely following recent announcements from OpenAI and Anthropic, underscores the rapid pace of innovation in the AI domain.
The model’s enhanced reasoning abilities have already garnered attention, with Gemini 3 achieving an impressive score of 37.4 on the Humanity’s Last Exam benchmark, surpassing previous benchmarks. Additionally, it has topped the LMArena leaderboard, showcasing its ability to drive user satisfaction.
Google has reported a substantial user base for the Gemini app, with over 650 million monthly active users and 13 million developers leveraging the model in their workflows. Alongside Gemini 3, Google introduced Google Antigravity, a coding interface powered by Gemini, offering a unique approach to agentic coding.
This release represents a significant advancement in AI technology, setting a new standard for reasoning depth and nuance. As Google continues to innovate in the AI space, Gemini 3 is poised to make an impact on various industries and applications.
Source: TechCrunch