Verified · Nov 16, 2026
Independently verifiedGemini 4 Hydrogen launches: Google's rapid-execution model reaches Search worldwide, the Gemini app, and AI Studio — flat per-million-token API pricing and a new xhigh thinking tier
4 sourcesGoogle's model post (November 12, 2026) introduced Gemini 4 Hydrogen: "Today, we’re introducing Gemini 4 Hydrogen, bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows." The family split per the post: "While our frontier models, Gemini 4 Carbon and Gemini 4 Argon, excel at deep, long-horizon reasoning on your hardest problems, Hydrogen is built for rapid execution across diverse agentic workflows." Availability per the post: "With Hydrogen, we’re also bringing 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." The Search post states: "Hydrogen is now available globally in Google Search for all users" and "This is the biggest model upgrade to Search yet." — Google's own superlative. Free-tier access is the post's own footnote: "free users will get access to Gemini 4 Hydrogen via auto model selection when it’s the appropriate model for the prompt". Pricing per the post: "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." 9to5Google (November 12) adds the frame: "After this week’s Carbon availability, Google today unveiled Gemini 4 Hydrogen instead of keeping the “Flash” branding." and "Compared to Gemini 3.8 Flash (which launched in September), Hydrogen is billed as a “major standalone upgrade” with “significant leaps in both speed and intelligence.”"
Why now
Hydrogen landed Thursday, November 12 and is, per the Search post, "now available globally in Google Search for all users" — creators who cover model picks therefore have a fresh tier to explain: Google's own four thinking levels ("low, medium, high (default), and a new xhigh tier"), the flat per-million-token API pricing, and the free-tier path via auto model selection. 9to5Google's frame — Carbon available earlier the same week, Hydrogen replacing the Flash branding — makes the Gemini 4 lineup naming itself a story worth explaining now.
Why it is worth publishing
Model-choice content is evergreen traffic for creator audiences, and this card carries the exact material a pick-guide needs: family positioning from the official posts, the API pricing pair, the free-tier footnote that most rollout coverage paraphrases loosely, and a media outlet's independent frame. The vendor-benchmark risk card keeps every number attributed while the creator still gets the whole map.
Evidence basis
Four opened sources, three official, all read in full this run: Google's model post (November 12, 2026; JSON-LD datePublished 2026-11-12T18:00:00+00:00 with dateModified 2026-11-13T18:27:54; 417,862 bytes raw HTML), the developers post (November 12; 380,260 bytes), the Search post (November 12; 367,535 bytes), and 9to5Google (November 12; 2026-11-12T18:01:41+00:00; 221,845 bytes). Heat rides the surface breadth — one model reaching Search globally, the Gemini app, AI Studio, and Antigravity in a single wave — and the naming story (Flash branding retired for Hydrogen per 9to5Google), tempered by the vendor-reported nature of every benchmark figure and the progressive rollout in the posts' own verbs.
“On Thursday, Google put a new model into Search worldwide and the Gemini app — Gemini 4 Hydrogen — and free users get it automatically when it's the right model for the prompt, per Google's own footnote.”
Angle
A model-pick explainer built on the official posts' own map: where Hydrogen sits against Carbon and Argon (rapid execution vs. deep long-horizon reasoning), what the four thinking levels and the xhigh tier change, what the flat $1.50/$7.50 per-million-token API pricing means for a creator's tooling budget, and how the free tier actually gets it — via auto model selection, per the post's own footnote.
Format
Long-form explainer
Demo idea
A lineup demo: run the same prompt against Hydrogen in the Gemini app and note which thinking level the model picker shows, then show the four thinking levels from Google's post ("low, medium, high (default), and a new xhigh tier"), then the Search side — ask one multi-part question and show the visual response elements Google lists ("more images, structured responses, comparison tables, and custom tools built for you") — closing card: API pricing pair, free-tier auto-selection line, and every benchmark number labeled as Google's own report.
Platform notes
Every benchmark and safety figure on this card is Google's own report about its own model — the AI Index reading (51 vs. 50, 54 at xhigh), the Orchestrator and 3.8 Flash comparisons, the partner-reported 2x-2.8x range, and the internal-evaluation safety percentages; attribute them to Google wherever reused, keep each "up to"/"more than"/"reaching" bound welded to its number, and never present "the biggest model upgrade to Search yet" as an independent verdict — it is Google's own superlative. Free-tier availability is exactly the footnote's scope (auto model selection "when it’s the appropriate model for the prompt"), not an on-demand picker guarantee; the rollout is progressive in the posts' own verbs ("rolling out globally", "is now available globally in Google Search"). The $1.50/$7.50 figures are API pricing, not a consumer plan price. The top thinking tier is Google's "xhigh" — 9to5Google's "Extra" is that outlet's rendering. No numbers merge across sources.
Usable claims
- Google's model post (published November 12, 2026; calendar-verified Thursday) states: "Today, we’re introducing Gemini 4 Hydrogen, bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows." and positions it inside the family: "While our frontier models, Gemini 4 Carbon and Gemini 4 Argon, excel at deep, long-horizon reasoning on your hardest problems, Hydrogen is built for rapid execution across diverse agentic workflows." The post's availability sentence reads: "With Hydrogen, we’re also bringing 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." Per the post: "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." and "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." The rollout sentence reads: "Starting today, Hydrogen is rolling out globally across our developer, enterprise, and consumer surfaces" — and the post's footnote gates the free tier: "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." The post states: "Hydrogen’s four thinking levels — low, medium, high (default), and a new xhigh tier — offer a wide range of options across quality, latency, and cost" and "To help make this multi-model collaboration automatic, we’re introducing Orchestrator in Google Antigravity. Rather than using a single model from start to finish, Orchestrator dynamically switches between models like Carbon and Hydrogen within the same session." — reporting that "On Terminal-Bench 4.0, Orchestrator reached near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster." On the 3.8 Flash migration the post states: "It delivers significant leaps in both speed and intelligence — 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." and "in early testing across complex graphs, SQL, and data-heavy enterprise workflows, our partners have seen Hydrogen run 2x to 2.8x faster while reducing token usage by 12% to 23%, compared to 3.8 Flash". The developers post (November 12) states: "Starting today, developers can build with Gemini 4 Hydrogen, bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows." and "Built for rapid iteration, Hydrogen handles repository coding and debugging, interactive web development, and multimodal workflows — running standalone across four configurable thinking levels or alongside frontier models like Gemini 4 Carbon and Gemini 4 Argon." and "With tighter instruction following and fewer fabricated claims, it helps you ship cleaner, production-ready code faster." The Search post (November 12) states: "Today, Google Search is bringing cutting-edge intelligence to everyone with Gemini 4 Hydrogen. This is the biggest model upgrade to Search yet." and "Hydrogen is now available globally in Google Search for all users, delivering more helpful, visual, and intuitive responses with improved model capabilities" — plus "Hydrogen’s advanced reasoning supercharges our query fan-out technique in Search, connecting you to a broader range of helpful links than before." and "You’ll see more images, structured responses, comparison tables, and custom tools built for you." and "No matter what language you ask in, you’ll now get more highly nuanced and helpful responses across all supported AI Mode languages." 9to5Google (November 12) adds the framing: "After this week’s Carbon availability, Google today unveiled Gemini 4 Hydrogen instead of keeping the “Flash” branding." and "Compared to Gemini 3.8 Flash (which launched in September), Hydrogen is billed as a “major standalone upgrade” with “significant leaps in both speed and intelligence.”" and reports "Hydrogen is the default model for AI Mode and is used for “your hardest questions” in AI Overviews."
Evidence pipeline
From the news
- Google blog: Introducing Gemini 4 Hydrogen (November 12, 2026)
- Google blog: Gemini 4 Hydrogen for developers (November 12, 2026)
- Google blog: Google Search gets a major upgrade with Gemini 4 Hydrogen (November 12, 2026)
- 9to5Google: Google announces Gemini 4 Hydrogen with 'significant leaps' in speed, intelligence (November 12, 2026)
Breakdown
Four opened sources, three official. The model post carries the launch mechanics — the family positioning against Carbon and Argon, the availability sweep (AI Studio, Gemini API, Gemini for business, the Gemini app, Google Search, Antigravity for paid subscribers), the AI Index reading (51 vs. 50, 54 at xhigh), the flat API pricing pair, the four thinking levels with the new xhigh tier, Orchestrator, and the free-tier footnote. The developers post carries the developer framing (repository coding, web development, multimodal workflows, "fewer fabricated claims") and the Search post carries the Search surface ("now available globally in Google Search for all users", query fan-out, visual responses, language coverage). 9to5Google adds what the official posts do not state: the Flash-branding observation, the "which launched in September" dating for 3.8 Flash, the model-picker 'Everyday help.' billing, the AI Mode default and AI Overviews detail, and the 'Extra' tier rendering (Google says xhigh). What stays vendor-reported: every benchmark and safety figure, several on Google's internal evaluations — none independently verified this run; the Search superlative is Google's own; free-tier access is footnote-scoped (auto model selection when appropriate); and the rollout stays in the posts' own progressive verbs.
Sources
Risks
- Before you present any Hydrogen number, open Google's own post and quote it with its attribution: say Google reports for the benchmark standings and in Google's internal evaluations for the safety figures; keep every up to, more than, and reaching bound attached to its figure; present free-tier availability exactly as the footnote states it (auto model selection when it is the appropriate model), label the $1.50/$7.50 pricing as API pricing, and when the top thinking tier comes up, use Google's own term xhigh — 9to5Google's Extra is that outlet's rendering.
Demo ideas
- A family-map card: Google's own split ("While our frontier models, Gemini 4 Carbon and Gemini 4 Argon, excel at deep, long-horizon reasoning on your hardest problems, Hydrogen is built for rapid execution across diverse agentic workflows.") with each row labeled by which official post states it, and the 9to5Google row carrying its own frame ("After this week’s Carbon availability, Google today unveiled Gemini 4 Hydrogen instead of keeping the “Flash” branding.").
- A thinking-levels explainer board: the four levels verbatim from the model post ("low, medium, high (default), and a new xhigh tier"), one row on what the post says each setting trades (quality, latency, cost), and a footer on the naming drift: Google says xhigh, 9to5Google renders it Extra.
- An availability-and-pricing card: the rollout sentence ("Starting today, Hydrogen is rolling out globally across our developer, enterprise, and consumer surfaces"), the Search availability line ("now available globally in Google Search for all users"), the free-tier footnote quoted in full, and the flat API pricing pair — each bound carried verbatim, no upgraded availability.