Google Gemini 3.1 Pro Launches With 77% ARC-AGI Score
Google launched Gemini 3.1 Pro on February 19, bringing the core reasoning breakthroughs from its experimental Deep Think model to everyday applications. The new model scores 77.1% on the ARC-AGI-2 benchmark—more than double the reasoning performance of Gemini 3 Pro—and is now rolling out across consumer and developer platforms. What Makes Gemini 3.1 Pro Different…
