Last week, Google released Gemini 3, its most powerful foundation model to date. Today, you can now find it even more easily with the new Gemini app and AI search interface. This newest version of the Gemini model comes in only seven months after the release of Gemini 2.5. It comes less than a week after OpenAI launched GPT 5.1. Given the current race between companies and countries to achieve all things AI, the timing of these releases couldn’t be more interesting.
Gemini 3 also performed remarkably well on Humanity’s Last Exam benchmark, scoring a 37.4, which is incredibly high. This score beats the former high score of 31.64 set by GPT-5 Pro. Since the Humanity’s Last Exam benchmark was specifically crafted to test for abstract reasoning and specialized knowledge, Gemini 3’s high level performance is all the more impressive. This very exciting enhancement further demonstrates Google’s continued focus on innovation with their generative AI technologies.
On top of that benchmark success, Gemini 3 placed first overall on LMArena. Based on an extensive human evaluation of user satisfaction, this external benchmark demonstrated Gemini 3’s innovation leadership. Gemini 3 has quickly become Google’s most powerful large language model (LLM) ever. This dual success holds the potential of better functionality and improved performance for users.
Tulsee Doshi, Google’s head of product for the Gemini model, is thrilled by the progress in reasoning capabilities. She’s convinced that this latest publication is a major step in that direction.
“With Gemini 3, we’re seeing this massive jump in reasoning.” – Tulsee Doshi
Among Gemini 3’s features, a new coding application aims to make it easier for users to interact with multiple tools. This is something that Koray Kavukcuoglu, CTO of DeepMind, touched on further, emphasizing the agent’s multi-purpose abilities to support users.
“The agent can work with your editor, across your terminal, across your browser to make sure that it helps you build that application in the best way possible.” – Koray Kavukcuoglu
Gemini 3, in many ways, represents a new epoch in AI development. Today, models are aimed squarely at improving user experience and providing real-world use cases. With major players such as Google and OpenAI rapidly advancing the capabilities of AI technology, users can expect to be in the forefront of technical development as these innovations come into play!
The pace Google is going through the AI innovation gauntlet is staggering. As industry watchers such as Russell Brandom have recently observed, these changes powerfully impact platform policy and new technologies that are on the frontier. With this release of Gemini 3, we see a major jump in technical capability. It also represents a more general strategic effort to increase consumer happiness in an extremely competitive industry.