Google releases Gemini 2.5 Pro update that supercharges its coding abilities

It’s been a little over a month since Google debuted Gemini 2.5 Pro. When this update arrived, Google claimed that the AI model excelled at a range of benchmarks, including science, math, and reasoning. Coding was also an area of focus, with the company boasting that the AI can create “visually compelling web apps and agentic code applications.” Ahead of Google I/O 2025, Gemini 2.5 Pro is now getting an update to supercharge its coding abilities.

Today, Google announced it is releasing an update to Gemini 2.5 Pro in early access. The company says it was planning on releasing the update at I/O 2025, but has decided to go live with it a little early due to “overwhelming enthusiasm.”

Dubbed Gemini 2.5 Pro Preview (I/O edition), this version of the AI model is said to have further improved coding and multimodal reasoning capabilities. Google highlights that this update makes the LLM particularly good at “building compelling interactive web apps.” Additionally, the tech giant says improvements have been made to “code transformation, code editing, and developing complex agentic workflows.”

For developers, this latest version of Gemini 2.5 Pro can be accessed through the Gemini API via Google AI Studio and Vertex AI. Meanwhile, it’s also available in the Gemini app for everyone else.