Although the Google I/O conference is still a few weeks away, the company has already begun unveiling some of its most notable new features. In response to positive user feedback, Google decided to move up the launch of a preliminary version of its new artificial intelligence model.
Gemini 2.5 Pro Preview (I/O Edition)
On Tuesday, Google announced early access to Gemini 2.5 Pro Preview (I/O Edition) , an enhanced version of its AI model that demonstrates significant improvements in programming. According to the company, the new model offers significant improvements in creating interactive web applications, as well as other software development tasks.
New capabilities include:
- Transforming existing code
- Editing and debugging code
- Creating workflows with intelligent agents
This update was originally scheduled to be officially unveiled at Google I/O, but the excitement generated by internal testing accelerated its public availability.
Performance and availability
Gemini 2.5 Pro has climbed the rankings, topping the WebDev Arena with an Elo score 147 points higher than its previous version. It also achieved 84.8% in the VideoMME benchmark , confirming its improvements in tasks related to audiovisual content.
The new version is now available to developers through the Gemini API in Google AI Studio and Vertex AI. The updated model can also be accessed by the general public through the Gemini app, where it powers tools like Canvas.
Next edition of Google I/O
Google I/O will take place on May 20 and 21 at the Shoreline Amphitheater in Mountain View, California. The main event, featuring the latest hardware and software developments, will take place on the first day at 10 a.m. (Paris time).
