Recent findings suggest that google made Gemini 4 Hydrogen the default model in AI Mode for all users worldwide, starting today. Sites that track how often AI Mode cites them now have a upcoming model to account for.

In an X thread announcing the change, Robby Stein, VP of Product for Google Search, said Hydrogen is also “being used for your hardest questions in AI Overviews.” Google’s Search post says the model is available globally in Search for all users.

Industry observers note that the switch came two days into the November core enhancement, which was still rolling out as of 2:05 p.m. ET today.

The Search post says Hydrogen is able to handle multi-part inquiries, inquiries that contain typos, highly specific inquiries, and follow commands more thoroughly than previous versions of Search.

According to recent reports, the post from Search says that Hydrogen also allows for better query fan-out. Query fan-out refers to how Search simultaneously executes multiple queries across each of the subtopics of an inquiry. According to the post, the result is what Google calls “connecting you to a broader range of helpful links than before,” and this includes websites which users may have never stumbled upon themselves. Stein’s thread says that the fan-out improvements result in showing more links to the web. Neither the post nor Stein’s thread provide counts for links or sites.

The post references more visual aids in answers through use of images, structured responses, comparison tables and additional customized tools available to users; additionally, the answers provided will be more nuanced in all languages supported by AI Mode.

The last default change I covered was at I/O back in May when Gemini 3.5 Flash became the global default for AI Mode for everyone. Neither the Search post nor Stein’s post stated what Hydrogen replaced.

Stein’s thread has Hydrogen handling the hardest AI Overview questions but doesn’t say which questions count as hardest or what answers the rest. The Search post doesn’t mention AI Overviews.

Google made Gemini 3 the default model for AI Overviews globally in January. At the time, I noted that Google had described looks to send complex questions to Gemini 3 and simpler ones to faster models, and hadn’t mentioned whether that would apply to AI Overviews.

Citation data you pull for AI Mode or AI Overviews this month now spans a model change on Nov. 12. The change applies to AI Mode’s default answers and to the hardest AI Overview questions.

Nov. 12 falls inside the core update window, which began Nov. 10. Logging both dates lets you compare your tracking data before and after each one, rather than reading this month as a single change.

Google’s claim that there are more links doesn’t include figures. You can check how it applies to your site by looking at the citation and referral numbers in your own reports.

Although the Search Status Dashboard shows that the November core revision is rolling out, its completion date will be logged as the second date, after Nov. 12.

You should watch for an announcement from Google on which AI Overview questions are classified as hardest so you can see how much of your AI Overview tracking will reflect the change.

→ Read original article