Google releases Gemini 3.1 Pro scoring 77.1% on ARC-AGI-2 benchmark
Technology
Google just dropped Gemini 3.1 Pro, its latest AI model, and it's a big leap from the last version.
This upgrade more than doubles its reasoning power, scoring 77.1% on the ARC-AGI-2 benchmark, and is the upgraded core intelligence that makes those breakthroughs possible.
Available in NotebookLM and Gemini app
Along with higher scores on key benchmarks, Gemini 3.1 Pro is available in NotebookLM and the Gemini app, so you'll see smarter help in more places soon.
Developers can try it out via AI Studio's API, while businesses get access through Vertex AI and Gemini Enterprise.
Just a heads-up: Google says its Deep Think mode is for heavier-weight science and engineering tasks.