Google has released Gemini 3.1 Pro, an upgraded core model designed to deliver stronger reasoning and complex problem-solving capabilities across its AI ecosystem. The model builds on the Gemini 3 series and is being rolled out in preview to developers, enterprises, and consumers.
Gemini 3.1 Pro follows last week’s update to Gemini 3 Deep Think and represents what Google describes as a step forward in foundational intelligence. The company said the model achieved a verified score of 77.1% on ARC-AGI-2, a benchmark that measures a system’s ability to solve novel logic patterns. According to Google, that result more than doubles the reasoning performance of Gemini 3 Pro.
The release underscores Google’s focus on strengthening core reasoning as competition intensifies among AI providers seeking to differentiate on intelligence rather than just scale.
Broad Platform Rollout
Gemini 3.1 Pro is now available in preview through the Gemini API in Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio for developers. Enterprise customers can access it via Vertex AI and Gemini Enterprise. For consumers, the model is rolling out in the Gemini app and NotebookLM, with higher usage limits for subscribers to Google AI Pro and Ultra plans.
The company said the model is intended for tasks that require advanced reasoning rather than simple question answering. Examples include synthesizing complex data sets, generating visual explanations, and supporting creative workflows.
One highlighted capability is code-based animation generation. Gemini 3.1 Pro can produce website-ready animated SVGs directly from text prompts. Because the output is generated as pure code rather than pixel-based video, files remain small and scalable without loss of quality. The feature reflects Google’s push to integrate generative AI directly into development workflows.
Toward More Agentic Systems
Google said the preview phase will help validate performance improvements before the model becomes generally available. The company plans further advancements in agentic workflows, where AI systems can coordinate multi-step tasks autonomously.
The release comes as AI firms race to improve reasoning benchmarks and enterprise integration. By embedding Gemini 3.1 Pro across consumer apps, developer tools, and enterprise platforms simultaneously, Google is signaling a strategy of unifying its AI stack around a stronger baseline model.
Since launching Gemini 3 Pro in November, Google said user feedback and rapid iteration have driven improvements. With 3.1 Pro, the company is positioning its AI offerings as more capable for complex scientific, research, and engineering challenges, while making the upgraded intelligence available in everyday applications.
The broader rollout reflects the growing importance of reasoning performance as AI adoption expands from chat-based interfaces into coding, content creation, and enterprise automation.
