Whether you’re an experienced developer or a vibe coder, Gemini 3 can help you bring any idea to life. Today, we’re excited to introduce Gemini 3, our most intelligent model built on state-of-the-art reasoning. Gemini 3 Pro delivers unparalleled results across every major AI benchmark compared to previous versions and excels at coding, mastering both agentic workflows and complex zero-shot tasks.
Gemini 3 Pro fits seamlessly into existing production agent and coding workflows while enabling new use cases previously not possible. It’s available in preview at $2/million input tokens and $12/million output tokens for prompts up to 200k tokens through the Gemini API in Google AI Studio and Vertex AI for enterprises. Additionally, you can utilize it via your favorite developer tools within the broader ecosystem and access it for free with rate limits in Google AI Studio.
Agentic Coding
Developers are increasingly creating software with AI assistance. Building on the momentum of Gemini 2.5 Pro, Gemini 3 Pro serves as a new foundation for what’s possible with agentic coding models. It scores 54.2% on Terminal-Bench 2.0, testing a model’s tool use ability to operate a computer via terminal.
Experience the power of this model in Google Antigravity, our new agentic development platform, along with Gemini CLI, Android Studio, and other coding products like Cursor, GitHub, JetBrains, Manus, and Cline.
Google Antigravity
To advance how models and IDEs work together, we’re introducing Google Antigravity to showcase Gemini 3’s capabilities. This agentic development platform enables developers to operate at a higher, task-oriented level by managing agents across workspaces while maintaining a familiar AI IDE experience. It’s a faster way to develop—you act as the architect, collaborating with intelligent agents that operate autonomously across the editor, terminal, and browser. These agents plan and execute complex software tasks, communicating their work via detailed artifacts. Download the public preview for free on MacOS, Windows, and Linux.
Gemini API
With Gemini 3, we’re releasing a client-side bash tool that empowers the model to propose shell commands for agentic workflows, including navigating local filesystems, driving development processes, and automating system operations. This is paired with a hosted server-side bash tool for multi-language code generation and secure prototyping. Additionally, Gemini hosted tools like Grounding with Google Search and URL context can now be combined with structured outputs, enhancing agentic use cases involving data fetching and extraction.
Vibe Coding
Gemini 3 Pro unlocks the true potential of “vibe coding,” where natural language is your only syntax. By improving complex instruction following and deep tool use, the model can translate high-level ideas into fully interactive apps with a single prompt. It handles multi-step planning and coding details, delivering richer visuals and deeper interactivity while you focus on creative vision. Gemini 3 Pro tops the WebDev Arena leaderboard with an impressive 1487 Elo.
Google AI Studio
Whether building a game with a single prompt, an interactive landing page from voice notes, or a full app from a napkin sketch, developers can bring ideas to life with Gemini 3. Google AI Studio is your fastest path from prompt to AI-native app. Build mode lets you add AI capabilities quickly, automatically wiring up the right models and APIs, while annotations enable fast iteration. Start building with Gemini 3 in Google AI Studio today.
Advanced Capabilities
Gemini 3 excels in multimodal understanding, setting new highs on MMMU-Pro for complex image reasoning and Video MMMU for video understanding. With a 1 million-token context window, developers see significant improvements in key multimodal use cases. You can now configure multimodal vision processing with granularity in the Gemini API for better control over latency and cost.
For visual reasoning, Gemini 3 Pro is best-in-class for document understanding, going beyond simple OCR to handle complex reasoning. Its spatial reasoning drives performance in embodied tasks like pointing and trajectory prediction, unlocking use cases in autonomous vehicles, XR devices, and robotics. The model also powers intelligent screen understanding for desktop, mobile, and OS screens.
In video reasoning, Gemini 3 Pro captures rapid action with high-frame-rate understanding and long-context recall, allowing for synthesizing narratives across hours of footage.
Build What’s Next, Today
Gemini 3 Pro is integrated into many developer products to fit your workflows and unlock new coding ways. Integrate it immediately via Google AI Studio and Vertex AI for Enterprise. We’re introducing a new thinking level and more granular media resolution parameters in the API, along with stricter validation for thought signatures to preserve model thoughts across multi-turn conversations.
Experience agentic capabilities by adding AI-native features to Android apps, automating workflows through Gemini CLI, or managing agents in Google Antigravity. For vibe coding, Google AI Studio’s Build mode lets you generate fully functional apps with a single prompt. Click “I’m feeling lucky” for inspiration, and let Gemini 3 Pro handle both creativity and code implementation.
The software landscape is shifting, and we’re committed to giving you the tools to push boundaries. This is just the start of the Gemini 3 era—we can’t wait to see what you build with Gemini 3 Pro!