Google puts Gemini 4 Hydrogen into Search's AI Mode and AI Overviews.

Google has made Gemini 4 Hydrogen, a new fast model, the default in AI Mode and is using it for the hardest questions in AI Overviews. It calls this the biggest model upgrade to Search yet, available to all users globally from 12 November.
Google lists four changes. Search should cope better with nuanced, multi-part or very specific questions, including ones with awkward typos, because the model follows instructions and calls tools more precisely. Its query fan-out, which splits one question into several searches run at the same time across related subtopics, is meant to surface a wider range of links than before. Answers should be more visual, with more pictures, comparison tables, structured layouts and custom-built tools. And Google says responses improve across all the languages AI Mode supports.
Hydrogen is the fast tier of Google's Gemini 4 family, alongside the larger frontier models Gemini 4 Carbon and Gemini 4 Argon, which Google positions for deep reasoning; Hydrogen is built for quick, multi-step agentic tasks. Google says it finishes tasks 2.5 times faster than Gemini 3.8 Flash at its default setting and scores 51 on the Artificial Analysis Intelligence Index, against 50 for Argon. Beyond Search, it is rolling out from today in the Gemini app, Google AI Studio, the Gemini API, Gemini for business and for paid users of Google's Antigravity coding tool. API pricing is $1.50 per million input tokens and $7.50 per million output tokens.
Google's posts give no figures on how much traffic the broader set of links sends to websites, and they say nothing about ads in AI Overviews or AI Mode. The claim that people will find sites they would have missed is Google's own.
Our read: publishers will judge that claim on their referral data. Fan-out rewards pages that answer the sub-questions behind a search, such as comparisons, specifications, prices and availability, not just the headline query. Brands in Saudi Arabia and the Gulf should check how their pages show up in AI answers in each language their customers search in, and watch organic referral traffic closely after the switch.
Why it matters
AI answers in Search are becoming more visual and link-heavy, which changes what content gets surfaced and clicked for brands that depend on organic search.
Related trends.

LVMH's media division signs a ChatGPT content and tools deal with OpenAI
LVMH's media division, which owns Les Echos, Le Parisien and Paris Match, has signed a partnership with OpenAI. Its journalism will appear in ChatGPT answers with links, and its staff get ChatGPT Enterprise.
Why it mattersDeals like this decide which publishers ChatGPT cites and links to, which shapes where brands get mentioned in AI answers.

Google: Gulf shoppers now spread peak season across 90 days
Google says the Black Friday weekend has become a 90-day shopping season in MENA, running from October to December. Its playbook, published as partner content in Campaign Middle East, cites app and search data from the UAE and Saudi Arabia.
Why it mattersGoogle's UAE and Saudi figures suggest peak demand builds in October and holds into December. The data is Google-commissioned, and the efficiency claims come from US studies.

Google tracking code now flags AI Max clicks on exact keywords
Google Ads has added a ValueTrack parameter, {search_term_matchtype}, that records whether a search reached an ad through the advertiser's own keyword or through AI Max. It closes a gap in which AI Max expansions carried the keyword's match type.
Why it mattersWithout it, AI Max expansions looked like ordinary exact or phrase matches in click data, hiding how much traffic Google's AI was adding beyond advertisers' keywords.