This Google Developers Blog post, dated May 6, 2025, announces the early release of Gemini 2.5 Pro Preview (I/O Edition). Authored by Logan Kilpatrick, it details the model's significantly enhanced coding capabilities, particularly in front-end/UI development, code transformation, editing, and creating sophisticated agentic workflows. Key highlights include its top ranking on the WebDev Arena leaderboard, state-of-the-art video understanding for "video to code" applications, and its ability to rapidly develop aesthetically pleasing and functional web apps. The update also incorporates improvements based on developer feedback, such as better function calling, and is available via Google AI Studio and Vertex AI.
Google Unveils Gemini 2.5 Pro Preview (I/O Edition) with Advanced Coding Features
Early Release Announced: Google released Gemini 2.5 Pro Preview (I/O Edition) on May 6, 2025, providing developers early access to its enhanced coding capabilities ahead of Google I/O.
Core Improvements: The update delivers significant improvements in coding, focusing on:
Front-end and UI development.
Fundamental coding tasks like code transformation and editing.
Creation of sophisticated agentic workflows.
Industry Acclaim & Performance Benchmarks:
Michele Catasta, President of Replit, lauded it as the "best frontier model when it comes to 'capability over latency' ratio."
It secured the #1 rank on the WebDev Arena leaderboard for its proficiency in building aesthetically pleasing and functional web applications.
Silas Alberti from Cognition's founding team noted its "leading performance on our junior-dev evals" and its ability to make correct judgment calls, likening it to a "more senior developer."
Key Capabilities and Highlighted Use Cases:
Video to Code: Features state-of-the-art video understanding (scoring 84.8% on the VideoMME benchmark), enabling new applications like the "Video to Learning App" in Google AI Studio, which creates interactive learning apps from YouTube videos.
Easier Feature Development: Strong at front-end web development, it helps developers implement new features by generating code that matches existing application styles (e.g., adding a video player to the Gemini 95 starter app).
Quick Concepts to Working Apps: Facilitates the rapid development of ideas into functional and visually appealing web applications, as demonstrated by a new dictation starter app featuring details like wavelength animations and responsive design.
Availability and Updates for Developers:
Accessible through the Gemini API in Google AI Studio for general developers and via Vertex AI for enterprise customers.
Existing Gemini 2.5 Pro users (version 03-25) are automatically transitioned to the new, improved version (05-06) with no action required.
The update addresses key developer feedback, including reducing errors in function calling and improving function calling trigger rates.
It remains available at the same price, and the model card has been updated to reflect the new version.
Call to Action: Google expresses excitement to see the innovative applications developers will build using the enhanced capabilities of Gemini 2.5 Pro.