Gemini 3.1 Pro release notes

TL;DR

Google DeepMind announced Gemini 3.1 Pro, an upgraded large‑language model that delivers markedly better reasoning on complex tasks and is now available in preview across the Gemini API, Vertex AI, the Gemini app, and Notebook LM.

What Gemini 3.1 Pro Is

Gemini 3.1 Pro is the newest iteration of Google’s Gemini 3 series, positioned as a “smarter baseline” for problems that require more than a simple answer. The model improves core reasoning capabilities and is engineered for tasks such as data synthesis, visual explanations, and creative coding.

Key Technical Improvements

  • Reasoning boost – On the ARC‑AGI‑2 benchmark, which tests novel logical patterns, Gemini 3.1 Pro achieved a verified score of 77.1 %, more than double the performance of its predecessor Gemini 3 Pro.
  • Advanced code generation – The model can emit production‑ready SVG animations, 3‑D interactive scenes, and full‑stack web interfaces directly from natural‑language prompts, preserving scalability and minimal file size.
  • Complex system synthesis – Demonstrated ability to integrate disparate APIs, exemplified by building a live aerospace dashboard that visualizes the International Space Station’s orbit from a public telemetry stream.
  • Creative reasoning – Translates literary themes into functional design, e.g., generating a personal portfolio website that captures the atmospheric tone of Wuthering Heights.

Availability Across Platforms

Audience Access Method
Developers (preview) Gemini API in Google AI Studio, Gemini CLI, the agentic development platform Google Antigravity, and Android Studio
Enterprises Vertex AI and Gemini Enterprise
Consumers Gemini app and Notebook LM (Pro & Ultra plans only)

All platforms now expose the upgraded core intelligence, allowing immediate experimentation with the model’s enhanced capabilities.

Demonstrated Use Cases

  • Code‑based animation – Generates crisp, scalable SVG animations from text prompts, enabling lightweight visual assets for web and mobile.
  • Interactive design – Produces immersive 3‑D simulations (e.g., a starling murmuration) with hand‑tracking interaction and generative audio that reacts to user input.
  • Data‑driven dashboards – Automates the creation of live telemetry visualizations, bridging complex API data to user‑friendly interfaces.
  • Literary‑inspired UI – Designs thematic web portfolios by reasoning about narrative tone and translating it into modern UI components.

Strategic Implications

  • Agentic workflows – By delivering stronger reasoning, Gemini 3.1 Pro accelerates the development of autonomous agents that can plan, execute, and iterate on multi‑step tasks without human supervision.
  • Enterprise AI adoption – The preview in Vertex AI and Gemini Enterprise signals Google’s intent to embed the model in large‑scale, production‑grade pipelines, potentially reshaping how companies automate analysis and decision‑making.
  • Creative coding democratization – High‑fidelity code generation lowers the barrier for designers and developers to prototype interactive experiences, expanding the market for AI‑augmented creative tools.

Roadmap and Future Releases

Google notes that Gemini 3 Pro launched in November 2025, and feedback from that rollout drove the rapid improvements seen in 3.1 Pro. The model is currently in preview; a general‑availability release is planned later in 2026 after further validation of agentic workflow capabilities.

How to Get Started

Developers can start experimenting by visiting the Gemini API preview page in Google AI Studio (model=gemini‑3.1‑pro‑preview) or by installing the Gemini CLI. Enterprise customers can enable Gemini 3.1 Pro in Vertex AI or Gemini Enterprise. Consumers on the Gemini app or Notebook LM with Pro/Ultra subscriptions will see higher usage limits and access to the new model automatically.

This article is based on the official Google DeepMind blog post “Gemini 3.1 Pro: A smarter model for your most complex tasks” published on 19 February 2026.

Sources