Gemini 3.1 Pro Tops ARC-AGI-2 Benchmark with 77.1% Score

Google DeepMind

Feb 19, 2026 · Updated Apr 25, 2026

Google released Gemini 3.1 Pro, scoring 77.1% on ARC-AGI-2 - more than double Gemini 3 Pro's score and ahead of Claude Opus 4.6. It's rolling out to the Gemini app, NotebookLM, and developer APIs including Gemini CLI and Antigravity.

Gemini 3.1 Pro is Google's upgraded reasoning model in the Gemini 3 series. On ARC-AGI-2 - a benchmark testing ability to solve entirely novel logic patterns - it scored 77.1% in Thinking (High) mode, more than double the 31.1% from Gemini 3 Pro and ahead of Claude Opus 4.6 (68.8%). The upgrade targets core reasoning rather than expanded multimodal features.

The practical difference shows in tasks where simple answers fall short: building a live aerospace dashboard from a public telemetry stream, generating interactive animated SVGs in pure code, or reasoning through literary tone to design a functional portfolio site. These are existing capabilities made more capable by stronger underlying reasoning.

3.1 Pro is in preview via the Gemini API in Google AI Studio, Gemini CLI, and Antigravity. Consumer access is rolling out through the Gemini app and NotebookLM on Pro and Ultra plans.

View the full update on blog.google

Google DeepMind

@GoogleDeepMindFeb 19

Gemini 3.1 Pro is here. We’ve significantly improved the model’s overall intelligence so it can solve tougher problems. 🧵

748

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

See all AI news & updates from Google →

Keep reading

Google Gemini 3.1 Flash Live Claims Top Spot for Production Voice Agents

Google's Gemini 3.1 Flash Live model reached the #1 position on the Tau Voice Bench leaderboard for real-time voice agents. The update delivers significantly lower latency and higher precision, signaling that multimodal voice AI is now reliable enough for production-grade applications.

Google Launches Gemini 3.1 Flash TTS With Natural Language Audio Tags

Google DeepMindApr 23

Google Launches Gemini 3.1 Flash TTS With Natural Language Audio Tags

Google released Gemini 3.1 Flash TTS, a text-to-speech model that uses natural language audio tags to control vocal style, pace, and delivery. This update allows users to direct AI speech like a human performance while maintaining low costs and high speed.

GoogleApr 24

Google Releases Prompting Formula for Granular Control of Gemini 3.1 TTS

Google released a formal prompting framework for Gemini 3.1 TTS that uses inline audio tags to control speech style and pacing. This update provides the specific syntax and constraints needed to direct AI voices like human actors, enabling dynamic and expressive vocal performances.

OpenRouter Adds Gemini 3.1 Flash Lite With 1M Context and Service Tiers

OpenRouterMay 7

OpenRouter Adds Gemini 3.1 Flash Lite With 1M Context and Service Tiers

OpenRouter integrated Google's Gemini 3.1 Flash Lite into its unified API, offering a 1 million token context window and multimodal processing for $0.25 per million input tokens. The update introduces a service tier parameter that allows developers to trade off latency for lower costs on high-volume agentic tasks.