Google has released the latest version of its powerful large language model, Gemini Pro. The company says version 3.1 is currently in preview and will be released to the public soon.
Google has released the latest version of its powerful large language model, Gemini Pro. The company says version 3.1 is currently available in preview and will be released to the public soon.
The company also shared statistics from independent benchmarks — including the Humanity's Last Exam test — which showed significantly better results for the model compared to its previous version.
Gemini 3.1 Pro also received praise from Brendan Foodie, CEO of AI startup Mercor, whose APEX testing system is designed to evaluate how well new AI models perform on real-world professional tasks.
“Gemini 3.1 Pro is currently at the top of the APEX-Agents rankings,” Foody said in a social media post, adding that the model’s impressive results demonstrate “how quickly agents are improving at performing real-world intelligent work.
"Building on the Gemini 3 series, the 3.1 Pro version represents a step forward in fundamental logical reasoning. The 3.1 Pro is a smarter and more powerful foundation for solving complex problems, as evidenced by our progress in rigorous benchmarks. In the ARC-AGI-2 test, which evaluates the model's ability to recognize completely new logical patterns, the 3.1 Pro demonstrated a verified result of 77.1%. This is more than double the logical reasoning performance of the 3 Pro model," the company's blog says .
Starting Thursday, Gemini 3.1 Pro is available in the Gemini app with expanded limits for Google AI Pro and Ultra plan users. 3.1 Pro is also now available in NotebookLM for Pro and Ultra subscribers. Developers and enterprises can already access the 3.1 Pro preview via the Gemini API in Google AI Studio, Antigravity, Vertex AI, Gemini Enterprise, Gemini CLI, and Android Studio.
Recall that last week , Google released an updated Gemini 3 Deep Think model for scientists, researchers, and engineers.
