Google unleashes gemini 3.1 pro: a cognitive leap in ai

Google has once again pushed the boundaries of artificial intelligence with the launch of Gemini 3.1 Pro, a cognitive leap forward in its quest to develop more intelligent and capable AI systems. This latest model boasts improved support for coding and a remarkable doubling of performance in competitor benchmark tests, outpacing GPT 5.2 from OpenAI.

A new era in ai reasoning

Gemini 3.1 Pro represents a significant advance in central reasoning, as it's designed to tackle complex tasks where a simple response isn't sufficient. The model's capabilities shine in the ARC-AGI 2 benchmark, which evaluates a system's ability to recognize and solve entirely new logical patterns. Gemini 3.1 Pro achieved a verified score of 77.1%, more than doubling the reasoning performance of its predecessor, 3 Pro.

Besting the competition

A head-to-head comparison with other top AI models demonstrates Gemini 3.1 Pro's supremacy in many popular benchmarks. The new model outperformed Gemini 3 Pro, Sonnet 4.6, Opus 4.6, GPT-5.3-Codex, and GPT-5.2 in tests like LiveCodeBenchPro, APEX-Agents, t2-bench, BrowseComp, MMMU Pro, MMMLU, MRCR v2 (8-needle), and Terminal-Bench 2.0, among others.

Humanity's last exam: a gauntlet for ai

In the challenging Humanity's Last Exam benchmark, which measures advanced reasoning capabilities, Gemini 3.1 Pro scored 44.4%, a slight improvement over its predecessor, Gemini 3 Pro, which achieved 37.5%. Notably, the new Google AI model outperformed OpenAI's GPT-5.2 by nearly 10%, with that model scoring 34.5%.

Availability and access

Gemini 3.1 Pro is now available for developers through various interfaces, including Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio. Businesses can access the model through Vertex AI and Gemini Enterprise, while consumers can experience its capabilities via the Gemini app and NotebookLM.

• Gemini 3.1 Pro is designed for complex, high-stakes tasks
• Boosts performance in competitor benchmark tests
• Improves coding support and central reasoning
• Available for developers, businesses, and consumers