Gemini 3.1 Pro: A smarter model for your most complex tasks

Google has released Gemini 3.1 Pro, an upgraded version of its core AI model, designed to handle more complex reasoning and problem-solving tasks. The model is now being rolled out in preview across Google's developer, enterprise, and consumer platforms, signaling a move to equip a broad user base with more powerful underlying intelligence for demanding applications.

The company claims a significant performance increase, citing a verified score of 77.1% on the ARC-AGI-2 benchmark, which measures a model's ability to solve novel logic problems. This score is more than double the reasoning performance of its predecessor, Gemini 3 Pro. Functionally, this translates to new capabilities such as generating animated SVGs directly from text prompts, building live data dashboards by synthesizing complex APIs, and creating interactive 3D visualizations from code. Gemini 3.1 Pro is accessible via the Gemini API, Vertex AI, Gemini Enterprise, the consumer Gemini app, and NotebookLM.

By releasing Gemini 3.1 Pro in a preview capacity, Google is strategically gathering data to validate the model's performance and refine its capabilities for more ambitious applications. The company explicitly mentioned a focus on advancing 'agentic workflows,' indicating that this release is a foundational step toward building more autonomous AI systems. This phased rollout, with tiered access for Pro and Ultra plan users on certain platforms, also reinforces Google's ongoing strategy to monetize its most advanced AI capabilities while building an ecosystem for developers and enterprises.

Google's preview release of Gemini 3.1 Pro is less about a single model upgrade and more about a public stress test for its agent-focused ambitions. The emphasis on improved reasoning and its integration across developer and enterprise tools suggests the company is actively building and validating the foundational intelligence required for the next wave of autonomous AI applications.