Google has unveiled an improved version of its flagship AI model, Gemini 2.5 Pro Preview (I/O Edition). According to the company, the updated version surpasses its predecessors in several important areas, including generating program code, creating web applications, and analyzing video content.
The release was timed just ahead of the annual Google I/O conference, where the company traditionally announces new technological advancements. This year is expected to bring a wave of AI-related premieres, including new models and products built on this technology. Gemini 2.5 Pro Preview is already accessible via the Gemini API and is also available on the Vertex AI and AI Studio platforms. Importantly, the cost of using the model remains unchanged from the previous version. As reported by TechCrunch, the updated model will also soon appear in the Gemini app for both web and mobile devices.
Key Features and Performance Enhancements
The main improvements in Gemini 2.5 Pro Preview (I/O Edition) include significantly enhanced capabilities for writing and editing code, along with improved development of complex agent workflows. According to Google, the model has achieved notable results in the creation of web applications and has topped the WebDev Arena Leaderboard, a rating that evaluates how effectively AI can create functional websites.
In the domain of video analysis, the model also delivered outstanding performance. It scored 84.8% on the VideoMME test, which is considered one of the industry benchmarks for AI capabilities in media content processing. These results suggest a wider range of practical applications for the model in complex media scenarios, notes NIX Solutions.
Developer Experience and Broader Usability
“For developers already using Gemini 2.5 Pro, the new version will not only improve the work with code, but also take into account key requests, including reducing the number of errors when calling functions,” Google noted in a blog post. The company also highlighted that the model now more effectively addresses aesthetic considerations when developing web interfaces, while maintaining manageability and flexibility.
This latest update represents a notable step forward in AI development and demonstrates Google’s continued focus on improving performance and user experience. Yet we’ll keep you updated as more integrations become available and additional features are rolled out.