Google's Gemini 3.1 Pro more than doubles reasoning, scores 77.1%
Technology
Google just rolled out Gemini 3.1 Pro, its latest AI upgrade that is now more than double the reasoning performance of the last version, scoring a strong 77.1% on a major benchmark test.
The original Gemini 3 was already beating rivals like Copilot, but this new release takes things up a notch.
Gemini 3.1 Pro HLE hits 44.4%
Gemini 3.1 Pro builds on recent advances in science, math, and coding skills, plus it is much better at handling everyday tasks: its Humanity's Last Exam (HLE) score jumped to 44.4%.
You can try it out right now on Google's AI platforms.
While Anthropic still leads some rankings, experts say Gemini's real power will show as more people put it to use in the real world.