This year’s Google I/O 2025 event happens on May 20th, and it’s probably going to focus so much on new Gemini AI features that Google intends to introduce in the near future that the company won’t have time for anything else. Android, once the obvious star of the show, is getting a special “Android Show” on May 13th, which will be streamed online. That’s a clear hint that I/O 2025 will be all about AI.
But Google couldn’t even wait for I/O 2025 to release a big update to its latest AI model, Gemini 2.5 Pro. The I/O Edition version has been testing so well with users that Google made it available to anyone to use.
It’s not surprising to see AI firms release new products whenever they are ready. It happens all the time, and we’re used to seeing AI announcements come out of the blue. I’m saying that because Google unveiled the first version of Gemini 2.5 Pro only a few weeks ago, making it available to Gemini users across the paid and free tiers.
That’s already a massive AI release. Gemini 2.5 Pro brings reasoning powers to Gemini, a long context window (1 million tokens), better coding abilities, and high benchmark scores. You wouldn’t necessarily expect an upgrade so soon, but Google still gave us one.
“We were going to release this update at Google I/O in a couple weeks, but based on the overwhelming enthusiasm for this model, we wanted to get it in your hands sooner so people can start building,” Google said in a blog post.
Google is calling it “early access,” and the model’s name is Gemini 2.5 Pro Preview (I/O Edition). Yeah, Google’s AI names aren’t that much better than OpenAI’s.
The focus of this upgrade is coding. Google says the updated version of 2.5 Pro “has significantly improved capabilities for coding, especially building compelling interactive web apps,” which already “builds on the overwhelmingly positive feedback to Gemini 2.5 Pro’s coding and multimodal reasoning capabilities.”
Google also has a blog on its Developers website that explains Gemini 2.5 Pro’s new coding abilities.
Also, Google has new benchmark scores to prove the new Gemini 2.5 Pro version is better than alternatives. Gemini 2.5 Pro Preview (I/O Edition) tops the WebDev Arena Leaderboard, which measures “human preference for a model’s ability to build aesthetically pleasing and functional web apps.” Google says the new model has a “state-of-the-art performance in video understanding, with a score of 84.8% on the VideoMME benchmark.”
Speaking of video, Google says the new Gemini 2.5 Pro model can create even better interactive apps based on a single YouTube video. The following video shows Gemini creating a fully functional quiz after “watching” a video (see clip above).
Developers who rely on the newest AI models to vibe-code apps will probably appreciate the new Gemini 2.5 Pro release the most. But regular Gemini users with access to the AI models should benefit from the improvements. Regardless of Google’s focus on coding, Gemini 2.5 Pro is the best AI model you can get from Google.
While Gemini 2.5 Pro is available for free to Gemini users, you’ll get limitations along the way, like rate limits and a shorter context window. If you’re subscribed to Gemini Advanced, you should have access to all of Google’s latest AI models, including Gemini 2.5 Pro Preview (I/O Edition).
Developers looking to experiment with the new model can use it in Google AI Studio and Vertex AI. They’ll want to check out Google’s more detailed blog at this link.
