Goggle just dropped Gemini 2.5 Pro Preview (I/O edition) featuring even stronger coding capabilities, for users to start building with ahead of Google I/O, this preview version brings notable enhancements in front-end and UI development, alongside stronger capabilities for core coding tasks like code transformation, editing, and building agentic workflows.

Leading in Front-End Web Development

Gemini 2.5 Pro now tops the WebDev Arena leaderboard, a benchmark that evaluates models based on their ability to create visually appealing and functional web applications. This achievement highlights the model’s strength in delivering high-quality front-end solutions. Companies like Replit and Cognition have integrated Gemini 2.5 Pro into their platforms, with Replit’s President, Michele Catasta, praising its reliability for latency-sensitive tasks, and Cognition’s Silas Alberti noting its ability to handle complex refactoring tasks akin to a senior developer.

Practical Applications for Developers

The updated model excels in several key areas:

1. Video-to-Code Workflows: With an 84.8% score on the VideoMME benchmark, Gemini 2.5 Pro combines strong video understanding with coding capabilities. For example, it can generate an interactive learning app from a YouTube video, complete with a polished UI, as demonstrated in Google AI Studio’s Video to Learning App.
2. Streamlined Feature Development: The model simplifies front-end tasks by generating features that align with existing design systems. For instance, it can add a video player to an app while matching style properties like colors, fonts, and spacing.
3. Rapid Prototyping: Gemini 2.5 Pro enables developers to quickly turn concepts into functional, aesthetically pleasing web apps. A standout example is the dictation starter app, which features responsive design, wavelength animations, and subtle hover effects, showcasing the model’s attention to detail.

Developers can access Gemini 2.5 Pro through the Gemini API in Google AI Studio , while enterprise users can leverage Vertex AI . This update maintainthe same pricing. The release also addresses developer feedback by improving function calling accuracy and trigger rates, with an updated model card providing full details.