Google has introduced Gemini 3.1 Pro, a smarter AI model that aims to boost advanced reasoning, enterprise productivity, and to simplify complex tasks.
The new development comes a week later when Google also released a major update to Gemini 3 Deep Think to solve modern challenges across various domains, including research, science and engineering.
Gemini 3.1 Pro, dubbed as the “smarter AI mode,” is meant to be deployed for those tasks or questions for which simple answers are not enough.
The new model, equipped with core intelligence and advanced reasoning abilities, will be helpful in decoding and performing difficult tasks than its predecessor, the Gemini 3 Pro.
The new version will be available for developers in preview via the Gemini API, for enterprises in Vertex AI and Gemini Enterprise. It is now available in the Gemini app and NotebookLM for Google AI Pro and Ultra subscribers.
Gemini 3.1 Pro is designed to take core reasoning to the next level by provisioning a smarter and more capable baseline for complex problem solving.
When it comes to scoring, the model achieved an ARC-AGI-2 score of 77.1 percent , which is more than double the reasoning performance of Gemini 3 Pro.
Moreover, users also benefit from its advanced reasoning capabilities in practical scenarios.
The model will also tackle the evolving problems related to scientific research. According to Google, “this model is the upgraded core intelligence that makes those breakthroughs possible.”
This version would also solve the research challenges, providing definite and simplified answers with clear guidelines.
Gemini 3.1 Pro has also outcompeted its predecessor in other benchmarks, scoring 44.4 percent on Humanity's Last Exam (HLE) and 77.1 percent on ARC-AGI-2 logic benchmark tests.
Taking to X, Brendan Foody, the CEO of AI startup Mercor, lauded the newly released Google’s AI model based on its performance.
“Gemini 3.1 Pro is now at the top of the APEX-Agents leader board,” Foody said, adding “how quickly agents are improving at real knowledge work.”