On Tuesday, Google released Gemini 3, its most advanced foundation model to date, marking a significant leap forward in large language model capabilities. The model is now accessible through the Gemini app and Google’s AI-powered search interface. Arriving just seven months after Gemini 2.5, this update strengthens Google’s position in the competitive AI landscape.
The launch timing places Gemini 3 in direct competition with OpenAI’s recently unveiled GPT 5.1 and Anthropic’s Sonnet 4.5. This rapid succession of model releases highlights the accelerating pace of AI development across the industry. Google’s engineers describe Gemini 3 as a dramatic evolution in multimodal reasoning and contextual comprehension, pushing the limits of what foundation models can achieve.
A more research-focused variant named Gemini 3 Deepthink will roll out to Google AI Ultra subscribers once it clears additional safety evaluations. This specialized model is designed for advanced problem-solving tasks requiring deeper reasoning capabilities. Google emphasizes that safety, transparency, and human alignment remain core priorities throughout this release cycle.
Tulsee Doshi, Google’s head of product for Gemini, said Gemini 3 demonstrates a massive leap in reasoning depth and contextual understanding. According to Doshi, the new model can handle multi-turn dialogue, generate richer analytical explanations, and respond with a level of nuance previously unseen. Early testers reported that Gemini 3 performs significantly better in factual consistency and long-context comprehension.
Performance data reinforces these claims. Gemini 3 scored 37.4 on the Humanity’s Last Exam benchmark, surpassing GPT-5 Pro’s top score of 31.64. It also ranked highest on LMArena, a human-led benchmark assessing real-world user satisfaction and interactive quality. These results highlight the model’s strength across cognitive reasoning, knowledge depth, and conversational precision.
As of November 2025, the Gemini app boasts over 650 million monthly active users. More than 13 million developers now integrate Gemini models into their daily workflows. Google reports a substantial rise in developer engagement as AI-driven tools become increasingly vital for software development, data analysis, and creative automation.
Alongside the model launch, Google introduced a new Gemini-powered coding environment called Google Antigravity. This tool features multi-pane agentic coding that allows developers to interact with AI through commands, code, and browsers simultaneously. Positioned as a direct rival to Warp and Cursor 2.0, Antigravity aims to redefine modern coding workflows by combining intuitive design with autonomous execution capabilitie
