Copilot Pay-As-You-Go Pricing: Ignite 2026 Verdict
Google shipped Gemini 4 Hydrogen on November 12, 2026, and buried in the launch post is a bigger decision than the model itself: Flash is gone. Not discontinued, not deprecated with a six-month sunset notice — just quietly absent from the naming going forward, replaced by a periodic-table codename that started with Argon and Carbon and now adds Hydrogen to the set. A brand Google has used since Gemini 1.5 Flash launched in May 2024 doesn’t get a mention in its own successor’s announcement.
That’s a strange way to retire a name most developers actually liked. It’s also probably the least important thing about this launch. We’ve covered Gemini 4 Argon and Gemini 4 Carbon individually, and both share one problem: almost nobody can actually use them yet. Hydrogen is the one most people will actually touch, because Google put it everywhere at once — including, for the first time with a new Gemini model, straight into the free tier on day one.
Quick Summary: What Happened
Detail Info Launched November 12, 2026, via Google’s official announcement Replaces Gemini 3.8 Flash — and retires the “Flash” brand name going forward Availability Gemini API, Google AI Studio, Gemini for Business, the Gemini app, and Google Search — free-tier users included from launch day Pricing $1.50 per million input tokens / $7.50 per million output tokens, flat across the full 1-million-token context window Speed claim 2.5x faster than Gemini 3.8 Flash at default thinking settings vs. Gemini 4 Argon Undercuts Argon’s $2/$10 introductory rate, which rises to $4/$20 after November 30, 2026 Bottom line: Google just gave away a faster, cheaper model to everyone with a Google account on the same day it quietly killed the brand name that built its reputation for cheap, fast AI.
Gemini 4 Hydrogen slots in as the efficiency tier of Google’s Gemini 4 family, sitting below the frontier-grade Argon and Carbon models on raw capability but far ahead of them on who’s allowed to use it. Coverage from 9to5Google, VentureBeat, and Android Headlines all landed the same day, confirming the model rolled out simultaneously across the Gemini API, Google AI Studio, Gemini for Business, the Gemini app, and Google Search — a wider same-day spread than either Argon or Carbon got.
The pricing is flat and simple: $1.50 per million input tokens, $7.50 per million output tokens, no step-up as you approach the full 1-million-token context window. Google’s own post frames the jump from Gemini 3.8 Flash less as an incremental refresh and more as a standalone upgrade — finishing tasks 2.5 times faster at the default thinking setting, more than tripling its score on Terminal-Bench 4.1 (57% versus 18.2%), and cutting fabricated claims by more than 5x. On the Artificial Analysis Intelligence Index, Hydrogen posts a 51 against Argon’s 50 — a model Google markets as its efficiency tier edging out the model it markets as frontier, at least on that one index, before Hydrogen even gets pushed to its highest thinking setting (54).
Google isn’t saying it in so many words — the announcement simply doesn’t use “Flash” anywhere — but the pattern is hard to miss once you line up the last two months of releases. BigGo Finance’s reporting put it plainly: Google retired the Flash brand with Hydrogen’s launch, folding its fast-and-cheap model line into the same periodic-table naming scheme it started with Argon and Carbon. Three models, three element names, zero mentions of the tier label that’s been on Google’s efficiency models for two and a half years.
It’s worth sitting with how long that name lasted. Gemini 1.5 Flash launched in May 2024 as Google’s answer to a simple problem: not every task needs a frontier model, and charging frontier prices for a customer-support chatbot or a document classifier was a bad trade for everyone except Google’s revenue line. Flash became shorthand across the industry for “the cheap, fast Gemini tier” — developers built pricing calculators around it, we’ve compared it directly against GPT-5.5 in prior coverage, and competitors shaped their own budget tiers partly in response to it. That’s not a small brand to walk away from without comment.
The practical reason is probably simpler than any rebrand strategy memo: Hydrogen isn’t really “Flash, but newer.” It’s being pitched, and benchmarked, as something that beats a model Google itself calls frontier on at least one widely cited index. Calling that “Flash” undersells it. Calling it “Hydrogen” puts it in the same naming family as Argon and Carbon, which signals parity it didn’t have under the old scheme — whether or not the model fully earns that signal on every workload.
This is the comparison that actually matters for anyone making a buying decision today, because Hydrogen is the first model in the Gemini 4 family a normal developer or business can simply use. Argon and Carbon both launched through Google’s restricted Fairwind access program for vetted cyber-defense partners, which we covered in detail when Argon shipped in September and again when Carbon followed in October. Hydrogen skips that line entirely.
On price, Hydrogen is the cheaper option by a wide margin. Argon’s introductory rate — $2 per million input tokens, $10 per million output — is already more expensive than Hydrogen’s $1.50/$7.50, and that introductory window closes on November 30, 2026, after which Argon’s standard rate doubles to $4 input / $20 output. Once that happens, Hydrogen will cost roughly a third of Argon’s input price and well under half its output price, for a model that, per Google’s own disclosed benchmark, edges Argon on the Artificial Analysis Intelligence Index.
That last point deserves a caveat the same way Argon’s own benchmark claims did when we covered them: one index isn’t the whole picture, and Google is grading its own launch. Argon still carries the frontier-tier positioning for complex enterprise and cybersecurity work, and nothing about Hydrogen’s launch changes the fact that Argon and Carbon remain gated behind Fairwind for most of the people who’d want to test that claim independently. What’s changed is that the cheap, available model just closed most of the gap with the expensive, restricted one — on paper, for now.
The free-tier rollout is the detail worth paying more attention to than the pricing table. Google didn’t hold Hydrogen back for paying subscribers first and trickle it down to free users later, the way most frontier launches work, including Google’s own. It went into the Gemini app’s automatic model selection for free users, into Search for everyone worldwide, and into AI Studio’s and the API’s free tiers, all on launch day. For a company whose last two Gemini 4 releases were available to a short list of vetted cyber-defense partners and nobody else, shipping the third one free to anyone with a Google account is a deliberate contrast, not an oversight.
That contrast tells you what Hydrogen is actually for. Argon and Carbon are Google’s answer to “can we claim the best model in the world.” Hydrogen is Google’s answer to “what does the other 99% of Gemini traffic — chatbots, Search AI Overviews, everyday coding assistance, the stuff that runs billions of queries a day — run on for the next year.” Pricing and distributing that model aggressively matters more to Google’s actual revenue and market position than another benchmark chart nobody outside Fairwind can verify.
Zoom out and this is the third Gemini 4 model in six weeks, after Argon on September 30 and Carbon on October 28. Google is shipping frontier-adjacent models at a pace neither OpenAI nor Anthropic has matched this cycle, and it’s doing it with a access strategy that splits cleanly down the middle: restrict the models good enough to worry about, give away the model built for everyday use. That’s a coherent strategy, even if the Flash-to-Hydrogen rename wasn’t explained as part of it.
It also puts real pressure on the rest of the “cheap frontier model” market. Our comparison of Gemini 3.5 Flash against GPT-5.5 from earlier this year treated Flash as the default-tier anchor for cost-sensitive production traffic. Hydrogen replaces that anchor at a lower price than Argon charges and with a same-day free rollout nothing else in the Gemini 4 family has gotten. Anyone running the kind of workload-weighted router we’ve recommended in our AI cost optimization coverage now has a new default to plug in on the Google side, and it’s worth re-running that math rather than assuming last quarter’s Flash pricing still applies.
I think the benchmark numbers are genuinely good for what Hydrogen is supposed to be — a fast, cheap, high-volume model, not a frontier flagship. Tripling a coding benchmark score and cutting fabrication rates by 5x over the prior generation are real, meaningful jumps, and Google deserves credit for shipping them to free users immediately instead of gating them behind a paid tier for a quarter first.
The naming decision bothers me more than it probably should, but it’s worth saying plainly: retiring a two-and-a-half-year-old brand without acknowledging it in the launch post reads like Google wants the comparison against the old Flash line to just stop happening. “Hydrogen beats Argon on one index” is a cleaner story than “the cheap tier finally caught up to what we were charging frontier prices for a year ago,” and only one of those framings makes it into the press release.
For buyers, the decision is close to automatic right now. If you’re running anything on Gemini 3.8 Flash, Hydrogen is a straight upgrade at the same price shape, and the free-tier access means you can test it today without a procurement conversation. If you were evaluating Argon for its benchmark lead, that lead just got thinner and more expensive to access than the model sitting one rung below it.
Gemini 4 Hydrogen is Google’s fast, low-cost Gemini 4 model, announced November 12, 2026. It replaces Gemini 3.8 Flash as Google’s efficiency-tier model, running 2.5x faster at default settings and launching free to Gemini app, Search, and AI Studio users on day one.
Yes. Google’s launch announcement for Hydrogen doesn’t mention “Flash” at all, following the same periodic-table naming convention it introduced with Gemini 4 Argon and Gemini 4 Carbon. The Flash name, in use since Gemini 1.5 Flash launched in May 2024, doesn’t appear going forward.
$1.50 per million input tokens and $7.50 per million output tokens, flat across the model’s full 1-million-token context window. Free-tier access is also available through the Gemini app, Google Search, and Google AI Studio.
Hydrogen costs less than Argon’s introductory rate of $2 input / $10 output per million tokens, and considerably less than the $4/$20 rate Argon moves to after its introductory pricing ends November 30, 2026. Google’s own benchmark data shows Hydrogen edging past Argon on the Artificial Analysis Intelligence Index, though Argon remains positioned as the frontier-tier model for complex enterprise and cybersecurity work, and access to it is still restricted through Google’s Fairwind Program.
Yes. Hydrogen rolled out to free-tier users in the Gemini app (via automatic model selection), in Google Search worldwide, and in Google AI Studio’s free tier on launch day — a wider free-tier rollout than either Gemini 4 Argon or Gemini 4 Carbon received.
The Gemini API, Google AI Studio, Gemini for Business, the Gemini app, Google Search, and for paid Google AI Pro and Ultra subscribers, Google Sheets and Google Antigravity.
Last updated: November 14, 2026. Sources: Google — Introducing Gemini 4 Hydrogen · 9to5Google — Google announces Gemini 4 Hydrogen · VentureBeat — Google launches Gemini 4 Hydrogen · Android Headlines — Gemini 4 Hydrogen replaces Flash · BigGo Finance — Google retires Flash brand with Gemini 4 Hydrogen · Mixed News — Gemini 4 Argon introductory price ends November 30 · Epoch AI — Gemini 1.5 Flash.
Related reading: Gemini 4 Argon vs GPT-6 Astra vs Claude Opus 5.5 · Gemini 4 Carbon: Top AI Model Nobody Can Use · Gemini 3.5 Flash vs GPT-5.5 · AI Cost Optimization Guide 2026 · Gemini Spark Is Live: What AI Pros Need to Know