Summary

Google has unveiled Gemini 3.1 Pro, a groundbreaking AI model that leads on 13 of 16 benchmarks and demonstrates unprecedented reasoning capabilities, particularly on the Arc AGI 2 benchmark. The model represents a strategic shift, with Google pricing it remarkably low despite being the richest tech company, signaling a fundamental change in how AI models should be evaluated. The key highlight is Gemini 3.1 Pro's dramatic improvement in novel reasoning, more than doubling its predecessor's score in just 90 days and achieving the largest single-generation reasoning gain in frontier model history. This development suggests that the future of AI is less about which specific tool you use and more about understanding the rapidly evolving landscape of AI problem-solving capabilities.