Gemini 3 Deep Think
Google's new AI just spotted a flaw in a physics paper that human reviewers missed.
Deep Dive
Google has released a major upgrade to Gemini 3 Deep Think, its specialized reasoning mode for science and engineering. It's now available in the Gemini app for AI Ultra subscribers and via API for select researchers. Early testers used it to find a logical flaw in a peer-reviewed math paper and optimize semiconductor crystal growth. The model also set a new standard of 48.4% on the 'Humanity's Last Exam' benchmark.
Why It Matters
This could dramatically accelerate scientific discovery and R&D by automating complex reasoning on messy, real-world data.