Google announced Gemini 4 Hydrogen for faster agent work

Google announced Gemini 4 Hydrogen, a Gemini 4 model it says is built for speed and everyday coding and agentic workflows, and said the model is rolling out in Search, the Gemini app, and developer and business tools. The company blog post, by Koray Kavukcuoglu, is dated November 12, 2026, with a publication timestamp of 6:00 p.m. UTC, which is 1:00 PM ET. This article summarizes that announcement and contemporaneous coverage by TechCrunch and The Verge. It is not a hands-on test of the model, and the performance figures below are company claims unless a separate outlet is quoted.

Real, reported: based on the public sources cited below. Nothing added or dramatized. Sources checked November 13, 2026, at 7:14 AM ET (EST). Benchmark scores, speed comparisons, pricing, and safety figures come from Google's post or from press accounts of that post. The opened sources do not present this release as a peer-reviewed paper. Those results should be read as a company claim, not as an independently verified benchmark.

What Google says it is releasing

The post begins: "Today, we're introducing Gemini 4 Hydrogen, bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows." It contrasts Hydrogen with what it calls its frontier models, Gemini 4 Carbon and Gemini 4 Argon, which it says "excel at deep, long-horizon reasoning on your hardest problems," while "Hydrogen is built for rapid execution across diverse agentic workflows."

Google said the launch also "bring[s] Gemini 4 to everyone — for all tiers of Google AI Studio, the Gemini API, Gemini for business, the Gemini app, Google Search, and paid Google AI Pro and Ultra subscribers in Google Antigravity." Later, the post says that in the Gemini app, "free users will get access to Gemini 4 Hydrogen via auto model selection when it's the appropriate model for the prompt," and that the model is available "in Google Search for all users." It also says that, starting today, Hydrogen is "rolling out globally across our developer, enterprise, and consumer surfaces."

On price and context length, the post states: "Gemini 4 Hydrogen is priced flat across its entire 1 million-token context window at $1.50 per million input tokens and $7.50 per million output tokens." It also describes four "thinking levels — low, medium, high (default), and a new xhigh tier." Those terms and prices are Google's.

Performance figures Google reported

The blog says Hydrogen "surpasses Argon on the Artificial Analysis Intelligence Index at its default high thinking level (51 vs. 50, and reaching 54 at its highest xhigh thinking setting), while often completing tasks faster than models in its class." Those index scores are the figures Google published. They are not scores we reproduced, and the opened post does not present them as an audit by Artificial Analysis that we independently checked.

Google also said Hydrogen is "a major standalone upgrade for teams migrating from Gemini 3.8 Flash," "finishing tasks 2.5 times faster at its high default setting while more than tripling Terminal-Bench 4.1 scores (57% vs. 18.2%), cutting fabricated claims by more than 5x, and adapting cleanly when a tool call fails or a search returns empty." It said that on "multi-step tasks like deep research, data analysis, and broad codebase navigation, Hydrogen ranks #5 on Agent Arena, arena.ai's live benchmark," and it listed other company-reported results, including "HLE-Diamond (58.8%)", "Vals Tax Agent Bench (52.3%)", "Vals Legal Research Bench (55.8%)", "Multimodal Knowledge Agent (MKA) benchmark (53.4%)", and "Agents' Last Exam v1.1 (71.6%)". Those percentages and rankings are Google's reported figures.

The post also introduces "Orchestrator in Google Antigravity," which it says "dynamically switches between models like Carbon and Hydrogen within the same session." "On Terminal-Bench 4.0," Google wrote, "Orchestrator reached near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster." That comparison is also a company claim.

In a section on safety, Google said Hydrogen "reduces overreach in agentic tasks by up to 81% and under-service by up to 79%, and cuts misaligned actions and task shortcutting across tool-use and coding evaluations by 12x." It also reported "a 96.7% honesty rate on our internal agent honesty evaluation," said Hydrogen "preserves 99.9% chain-of-thought legibility with zero deceptive reasoning edits across more than 200,000 audited turns," and said it "is more than 2.3x more resilient on the Gray Swan IPI benchmark" against indirect prompt injection. Those safety figures are Google's internal or vendor-benchmark claims, not independent test results published in the materials we opened.

What the announcement says about Search

Google wrote that Hydrogen "is significantly upgrading Search, delivering more helpful, intuitive responses for all users globally." It said people can ask "in your natural flow of thought — even if you have multiple parts to your query, hard-to-understand typos, or highly specific requests," that responses will be "highly adaptive, visual, and easy-to-scan," and that the company is "upgrading our query fan-out technique in Search, connecting you to a broader range of helpful links than before."

TechCrunch, in an article by Sarah Perez time-stamped 11:27 AM PT on November 12, which is 2:27 PM ET, reported the Search rollout and wrote that the updates are coming "via the company's newly released Gemini 4 Hydrogen model, which Google claims is its 'biggest model improvement to Search yet.'" TechCrunch also repeated Google's comparison that Hydrogen completes tasks 2.5 times faster than Gemini 3.8 Flash and that, at its default high thinking level, it beats Gemini 4 Argon on the Artificial Analysis Intelligence Index. Those repeated figures remain company claims.

The Verge, in an article by Stevie Bonifield time-stamped 4:55 PM ET on November 12, likewise summarized the official feature list and quoted Google's Robby Stein, identified as VP of product at Google Search, as saying Gemini 4 Hydrogen is "the biggest model upgrade to Search yet" and "now the new default model in AI Mode & being used for your hardest questions in AI Overviews." Those Stein remarks appear in The Verge's account. The opened Google blog post describes the Search changes in Google's own words, quoted above, and does not include those Stein sentences in the passages we cite.

The Verge also wrote that, according to Google, the model "matches or outperforms some of the competing models in its class," like Claude Sonnet 5.5 and GPT-6.1 Terra. Competitive comparisons of that kind should be read as Google's positioning, not as a joint evaluation by those other labs.

How to read the claims

Google is announcing a production model and a broad rollout, not publishing a journal paper in the materials we opened. Speedups, benchmark scores, safety percentages, and pricing should be treated as the company's reported results. Where TechCrunch or The Verge repeat those figures, they are still repeating Google's claims unless the outlet describes a test of its own. The opened articles do not.

Our advice (general guidance)

If you use Google Search, the Gemini app, or the Gemini API, treat this as a product change to check in the official product or documentation you already use, not as a guarantee that answers will be accurate. Company benchmark scores are not a substitute for trying the model on your own tasks, especially if you rely on it for code, travel, medical, or financial questions. Keep a human review step for anything that will be published, spent, or sent to someone else.

These are editorial precautions, not individualized legal, financial or security advice.

Sources

Published November 13, 2026 by FraudMeNot Team. Label: Real, reported.

Disclaimer. General education only, not legal, financial or security advice. No guarantee. Consult a qualified professional. News summaries are based on the sources listed and may change as more is reported.

General education only, not legal, financial or security advice. Not affiliated with the organizations discussed. No sponsorship or affiliate links. Information may change as further evidence becomes available.