Google has incorporated Gemini 4 Hydrogen into its search functionality, and users around the world can start using it today. Officials have described this as the largest AI upgrade in the history of search engines.
The most noticeable change is in the way questions are formulated. In the past, one had to break down keywords separately when searching for information; now, it’s possible to combine several questions into a single sentence. Even if there are spelling mistakes or if the requirements are detailed and specific, that doesn’t matter anymore.
When searching, the system breaks down the issue into several sub-topics and searches for information related to each of them. As a result, the websites cited in the answers will be more numerous and diverse than before. The answers themselves also become more “visual”: they include images, comparison tables, and sometimes even small tools that users can click on directly.
The Hydrogen model within the Gemini series is specifically designed to focus on speed. According to data provided by Google, it completes tasks 2.5 times faster than the previous generation, the Gemini 3.8 Flash. In a comprehensive intelligence ranking, it scored 51 points, which is even one point higher than the larger Gemini 4 Argon model.
Developer price: Entering “$1.5 per million tokens” results in the output “$7.5”.
For those who work in web development, this upgrade is even more noteworthy than the model itself. The fact that search queries are now broken down into multiple sub-topics and processed simultaneously means that pages that don’t appear on the first page still have a chance of being shown; however, since users can directly see the tables and conclusions on the results page, there is one less reason for them to visit the original website.
It is currently not possible to use Google Search within China, but similar changes will inevitably be implemented in various AI search engines sooner or later.