Overview
Gemini 3.1 Pro is Google's latest update to the Gemini 3 series, building directly on the foundation of Gemini 3 Pro with a significant step forward in core reasoning capabilities. Released on February 19, 2026, this preview version targets developers, enterprises, and consumers who need AI that can handle complex, multi-step problems — not just simple queries.
The headline benchmark result is a 77.1% score on ARC-AGI-2, more than doubling Gemini 3 Pro's performance on the same test. Beyond benchmarks, 3.1 Pro is designed for practical applications: generating animated code, synthesizing data from complex APIs, and building interactive experiences from natural language prompts. It is currently rolling out in preview across the Gemini app, NotebookLM, Vertex AI, and developer tools.
What's New
Significantly Improved Reasoning
Gemini 3.1 Pro achieves a verified score of 77.1% on ARC-AGI-2, a benchmark that evaluates a model's ability to solve entirely new logic patterns it has never encountered before. This represents more than double the reasoning performance of Gemini 3 Pro on the same evaluation. For users, this translates to more reliable performance on tasks that require multi-step inference, complex data synthesis, and novel problem-solving rather than pattern recall.
Code-Based Animation and SVG Generation
3.1 Pro can generate website-ready, animated SVGs directly from text prompts. Because outputs are built in pure code rather than pixels, they remain crisp at any scale and maintain small file sizes compared to traditional video assets. This makes 3.1 Pro particularly useful for developers and designers who need scalable, lightweight visual assets without manual design work.
Complex System Synthesis
The model demonstrates improved ability to bridge complex APIs and user-friendly interfaces. In Google's demonstration, 3.1 Pro built a live aerospace dashboard by successfully configuring a public telemetry stream to visualize the International Space Station's orbit — entirely from a natural language prompt. This capability is relevant for engineers and data teams who need to turn raw API data into interactive dashboards without extensive manual coding.
Interactive Design and Creative Coding
3.1 Pro can build multi-sensory interactive experiences from descriptions. In one example, the model coded a complex 3D starling murmuration with hand-tracking controls and a generative audio score that adapts to the flock's movement. In another, it translated the atmospheric tone of Wuthering Heights into a complete, styled personal portfolio website — reasoning through literary themes to produce a functional design rather than summarizing the source material.
Availability & Access
Gemini 3.1 Pro is currently in preview and rolling out in phases:
- Consumers: Available in the Gemini app with higher usage limits for Google AI Pro and Ultra plan subscribers. Also available on NotebookLM for Pro and Ultra users.
- Enterprises: Available in preview on Vertex AI and Gemini Enterprise for enterprise testing and evaluation.
- Developers: Available in preview via the Gemini API through Google AI Studio, Gemini CLI, Google Antigravity (agentic development platform), and Android Studio.
Google says it will validate 3.1 Pro in preview before making it generally available, with further work planned on agentic workflows. No GA date has been confirmed.
Pricing & Plans
Gemini 3.1 Pro pricing operates across three separate tracks: subscription plans (Google AI app), API token billing (developers), and custom contracts (enterprise). The table below covers subscription tiers only.
| Plan | Price | Access to 3.1 Pro |
|---|---|---|
| Google AI Free | Free | Limited access (availability may vary by region/account) |
| Google AI Pro | $19.99/month | ✅ Included (higher limits) |
| Google AI Ultra | $249.99/month | ✅ Included (highest limits) |
| Gemini API (Developer) | Pay-per-token | ✅ Preview access |
| Vertex AI / Enterprise | Custom | ✅ Enterprise access |
Pricing tiers apply differently across access points: the $19.99/month figure refers to the Google AI Pro subscription; the Gemini API is billed separately per token (see Gemini API pricing); Vertex AI and Enterprise pricing is custom contract. Free-tier users may have limited access to 3.1 Pro in the Gemini app — Pro and Ultra plans provide higher and more stable limits.
Best For
- Researchers and engineers who need a model that can tackle novel logic problems and synthesize data from complex or undocumented APIs
- Frontend developers and designers looking to generate scalable animated SVG assets and interactive UI components from natural language descriptions
- Data teams who want to convert raw API data or telemetry streams into live dashboards without extensive manual coding
- Enterprise teams already on Google Cloud who want to test the latest Gemini reasoning capabilities through Vertex AI before general availability
- Developers building agentic workflows who want to validate 3.1 Pro's capabilities in preview via Google AI Studio or Gemini CLI
FAQ
How does Gemini 3.1 Pro compare to Gemini 3 Pro?
Gemini 3.1 Pro achieves 77.1% on ARC-AGI-2, which is more than double the score of Gemini 3 Pro on the same benchmark. Both models share the Gemini 3 architecture, but 3.1 Pro represents an upgraded core intelligence with improved reasoning depth and code generation quality.
Is Gemini 3.1 Pro available to free users?
Free users may have limited access to 3.1 Pro in the Gemini app, though availability varies by region and account. Google AI Pro ($19.99/month) and Ultra ($249.99/month) plans provide higher and more stable limits. Developers can access 3.1 Pro in preview via the Gemini API with pay-per-token pricing.
Is 3.1 Pro production-ready?
Not yet. Google is releasing 3.1 Pro in preview to validate the update and continue work on agentic workflow improvements before general availability. It is available in Vertex AI and Gemini Enterprise for enterprise testing, but Google has not confirmed a GA date.
What platforms support Gemini 3.1 Pro?
Gemini 3.1 Pro is available on the Gemini app, NotebookLM, Vertex AI, Gemini Enterprise, Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio. Consumer access requires a Pro or Ultra plan; developer access is via the Gemini API in preview.
What is ARC-AGI-2 and why does the score matter?
ARC-AGI-2 is a benchmark designed to measure a model's ability to solve entirely new logic patterns — tasks it has never been trained on. Unlike benchmarks based on knowledge recall, ARC-AGI-2 tests genuine reasoning ability. A 77.1% score indicates that 3.1 Pro can handle novel, multi-step logical problems significantly better than its predecessor, making it more useful for complex, real-world tasks beyond standard Q&A.
