Updated Gemini 2.5 Pro is New Coding Champ
- By Paul Mah
- May 15, 2025
Citing overwhelmingly positive feedback, Google has released the highly anticipated update of its Google Gemini 2.5 Pro ahead of the company’s annual Google I/O developer conference, scheduled for later this month.
With the latest improvements, Google said the Gemini 2.5 Pro (I/O Version) now leads the WebDev Arena Leaderboard, exceeding the score of its previous version by over 100 points. This leaderboard measures human preference for a model’s ability to build aesthetically pleasing and functional web apps.
The previous version of Gemini 2.5 Pro was released in March and introduced chain-of-thought reasoning. Even then, it outperformed other AI models from OpenAI and Anthropic in coding benchmarks. The latest update further enhances its capabilities, particularly in front-end web development.
As explained by TechRepublic here, developers can now easily automate tasks like matching colors, fonts, margins, and more whenever new features are added to their web projects. In fact, the dictation starter app built with Gemini 2.5 Pro removes all the technical work of building a user interface (UI) by providing default values.
“Beyond UI-focused development, these improvements extend to other coding tasks such as code transformation, code editing, and developing complex agentic workflows,” said Tulsee Doshi, senior director of product management at Google, in a blog post.
According to Silas Alberti, a founding team member at Cognition AI – the company behind Devin – the updated Gemini 2.5 Pro achieves “leading performance” on junior-dev evals.
“It was the first-ever model that solved one of our evals involving a large refactor of a request routing backend. It felt like a more senior developer because it was able to make correct judgment calls and choose good abstractions,” said Alberti.
The latest Gemini 2.5 Pro Preview is now available to developers via Google AI Studio and Vertex AI. It is also accessible to users through the Gemini app.
Image credit: iStock/Google
Paul Mah
Paul Mah is the editor of DSAITrends, where he report on the latest developments in data science and AI. A former system administrator, programmer, and IT lecturer, he enjoys writing both code and prose.