Is Gemini 4 Hydrogen free?

In most places, yes. Free Gemini app accounts get it “via auto model selection when it’s the appropriate model for the prompt,” Search uses it for everyone, and the Gemini API’s free tier includes it, rate-limited and with your content “used to improve our products.” Google AI Plus, at $4.99 a month, lists “Access to Gemini 4 Hydrogen” in the app.

Can I use Gemini 4 Hydrogen in Antigravity for free?

Google’s pages disagree. Its announcement and Antigravity’s blog say Hydrogen is “rolling out across paid Google AI Pro and Ultra subscribers,” but Antigravity’s models page marks it for Free and Google AI Plus too, and its plans page lists it among the “baseline features” of “All plans.” Check your model picker; Google hasn’t explained the difference.

How much does Gemini 4 Hydrogen cost?

On the Gemini API’s paid tier, $1.50 per million input tokens and $7.50 per million output tokens, thinking included, flat across its 1 million-token window, and $0.075 for cached input. That’s twice Gemini 3.8 Flash’s price through December 31, 2026, the same from January 1, and below Argon’s at every date.

Is Gemini 4 Hydrogen better than Gemini 4 Argon?

On Artificial Analysis’s Intelligence Index, slightly: 51 to Argon’s 50 at each model’s high setting, and 54 at Hydrogen’s xhigh. Its tasks cost $1.87 there against Argon’s $9.95, and took 6 minutes against 30. Google still puts Argon and Carbon ahead for “deep, long-horizon reasoning on your hardest problems.”

What is Orchestrator in Antigravity?

A choice in Antigravity’s model picker that pairs two models in one conversation, with Gemini 4 Carbon planning and reviewing and Hydrogen doing the edits, builds and tests. Google says it reached “near-Carbon task success rates” with 24% less quota and 19% faster. Antigravity’s models page lists it for Google AI Ultra only, and you can’t set its thinking level.

Should I switch from Gemini 3.8 Flash to Hydrogen on the API?

Google recommends it, calling Hydrogen the “Recommended upgrade from Gemini 3.8 Flash and earlier Gemini 3 and 2.5 models.” Budget for twice 3.8 Flash’s token price until January 1, 2027, and check the migration list first. Hydrogen defaults to `high` thinking, rejects `minimal`, ignores `temperature` and doesn’t support explicit caching.

Inter-office memo

To: Everyone using Gemini or Antigravity
From: Mo, Chief Executive
Date: November 12, 2026, 3:10 PM CST
Re: Gemini 4 Hydrogen

Gemini 4 Hydrogen is out, free in the Gemini app and API. Who gets it in Antigravity

Google released Gemini 4 Hydrogen on November 12, 2026, a fast model for everyday coding and agent work that edges out Gemini 4 Argon on Artificial Analysis’s index for a fifth of the cost per task. It’s free in the Gemini app and Search and on the API’s free tier, and $1.50 in and $7.50 out per million tokens paid. Google says Antigravity gets it on paid Pro and Ultra, though its models page marks every plan.

What Google released on November 12

On November 12, 2026, Google introduced Gemini 4 Hydrogen, “bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows.” It’s the third Gemini 4 model, after Argon on October 13 and Carbon, which reached the public two days earlier. Google keeps those two for “deep, long-horizon reasoning on your hardest problems,” and says Hydrogen “is built for rapid execution across diverse agentic workflows,” on its own or as the fast half of a team with Carbon.

It is also the first Gemini 4 model on free accounts. Google’s developer guide calls it the “Recommended upgrade from Gemini 3.8 Flash and earlier Gemini 3 and 2.5 models,” and its model ID is gemini-4-hydrogen. Input, thinking and output share one 1,048,576-token window, and it has four thinking levels: low, medium, high, the default, and a new xhigh.

Who gets it, and where

Here is Hydrogen, plan by plan, as Google’s pages read on November 12:

Plan or account	Price	Gemini app	Antigravity	Orchestrator
Free	$0	By automatic model selection	On the models page, not in the announcement	No
Google AI Plus	$4.99 a month	Yes	On the models page, not in the announcement	No
Google AI Pro	$19.99 a month	Yes	Yes, on a paid plan	No
Google AI Ultra	$99.99 or $199.99 a month	“Highest access”	Yes	Yes
Gemini Enterprise	Per seat	Gemini for business	Yes	Not on the models page

Developers get it on the Gemini API’s free tier, rate-limited, and its paid tier, and in AI Studio and Android Studio. The app column is from Google’s AI plans and the announcement, whose footnote says “free users will get access to Gemini 4 Hydrogen via auto model selection when it’s the appropriate model for the prompt.” The free plan’s card still lists “Access to our Flash-Lite model.” Search uses Hydrogen for everyone: it’s “now available globally in Google Search for all users,” says Google’s Search post. Gemini in Google Sheets starts with Pro and Ultra.

Antigravity is where Google’s pages disagree. The announcement brings Hydrogen to “paid Google AI Pro and Ultra subscribers in Google Antigravity,” and Antigravity’s own post says “Starting today, Gemini 4 Hydrogen is rolling out across paid Google AI Pro and Ultra subscribers.” But Antigravity’s models page marks Hydrogen in every column, Free and Google AI Plus included, and its plans page now lists “Use of Gemini models, including Gemini 4 Hydrogen, Gemini 3.8 Flash” among the “baseline features” that “All plans receive.” Google hasn’t said which is right, or whether free accounts get it later in the rollout. On a free account, look in your model picker before you count on it.

In the Antigravity CLI, --model gemini-4-hydrogen-xhigh or --effort xhigh picks the deepest level, which the app labels Extra. Neither GitHub’s nor Cursor’s changelog had added Hydrogen by the afternoon of November 12.

What it costs on the API

Google prices Hydrogen “flat across its entire 1 million-token context window.” Here it is beside the models it sits between, from the API’s pricing page, in US dollars per million tokens:

Model	Free tier	Input	Output, thinking included	Cached input
Gemini 4 Hydrogen	Yes	$1.50	$7.50	$0.075
Gemini 3.8 Flash, through December 31	Yes	$0.75	$3.75	$0.075
Gemini 3.8 Flash, from January 1, 2027	Yes	$1.50	$7.50	$0.15
Gemini 4 Argon or Carbon, through November 30	No	$2	$10	$0.10
Gemini 4 Argon or Carbon, from December 1	No	$4	$20	$0.20

So for now Hydrogen costs twice what 3.8 Flash does per token, and from January 1 the two cost the same. The free tier comes with a condition: content sent there is “Used to improve our products,” and on the paid tier it isn’t. In Antigravity, Hydrogen draws on your plan’s quota, which Google doesn’t size.

How it compares

Google 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).” Against 3.8 Flash, it claims tasks finish “2.5 times faster,” Terminal-Bench 4.1 scores more than triple, “57% vs. 18.2%,” and fabricated claims fall by “more than 5x.”

Artificial Analysis’s own leaderboard has the same index scores. Here are Hydrogen and its neighbors on November 12, with what a task cost Artificial Analysis at API prices:

Model	Index	Cost per task	Time per task
Gemini 4 Carbon (high)	60	$7.76	27.1 min
Claude Sonnet 5.5 (max)	55	$10.02	33.0 min
GPT-6.1 Sol (max)	55	$1.27	22.1 min
Gemini 4 Hydrogen (xhigh)	54	$3.16	13.0 min
Claude Opus 5.5 (high)	52	$3.46	9.5 min
Gemini 4 Hydrogen (high)	51	$1.87	6.0 min
Gemini 4 Argon (high)	50	$9.95	30.2 min
Gemini 3.8 Flash (high)	31	$7.77	14.8 min

At its default, Hydrogen scores a point above Argon for about a fifth of the cost and time per task. It isn’t the cheapest model at its level: GPT-6.1 Sol at max effort scores higher for less. And its weak spot is terminal work. The Decoder reports that Artificial Analysis’s testing put Hydrogen at 54% on Terminal-Bench 4.0, against 64% for Claude Sonnet 5.5 and 56% for GPT-6.1 Sol, each at max.

Orchestrator, for Ultra

Hydrogen arrived in Antigravity with Orchestrator, a new entry in the model picker. It isn’t a router, the models page says: “Orchestrator isn’t a router that classifies your prompt upfront and sends the whole request to a single model.” Each message starts with Carbon in “a read-only planning phase”; Hydrogen then edits files, runs builds and tests, and hands control back “If Gemini 4 Hydrogen runs into friction, such as a verification step that fails twice.” Carbon reviews progress on longer tasks.

Google says it reached “near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster,” on Terminal-Bench (its main post names version 4.0, Antigravity’s 4.1). The models page calls it the “recommended default” for “Most day-to-day development.” Two limits: it “runs with a fixed configuration, so you can’t choose a thinking level for it,” and it’s “rolling out for Google AI Ultra subscribers, with additional models and plans supported in the future.” The models page lists it for Ultra alone, though Antigravity’s post adds that “Enterprises can access both in Gemini for business.”

What to do now

On a free or AI Plus account, check Antigravity’s model picker. If Hydrogen is there, try it on a real task in place of Gemini 3.8 Flash. If it isn’t, two of Google’s own routes are free: the Gemini app, and Antigravity CLI with a free Gemini API key, where “Model requests go directly to the Gemini API” and Hydrogen is on the free tier.
On Google AI Pro, make Hydrogen your everyday model. Google suggests it for “package upgrades, code deduplication, scaffolding boilerplate, writing unit tests for well-defined functions,” and Argon, a point lower on Artificial Analysis’s index, took five times as long per task in its tests.
On Ultra, try Orchestrator first, and pick Carbon directly when “every intermediate step needs maximum reasoning depth.”
On the API, read the migration list before you switch. Change the ID from gemini-3.8-flash to gemini-4-hydrogen, raise any low max_output_tokens cap to 65,536, and remove temperature, top_p, top_k and explicit caching. Hydrogen “rejects minimal with an error.” Google’s managed Antigravity agent now uses Hydrogen by default.

What Google hasn’t said

Whether free and AI Plus accounts have Hydrogen in Antigravity. The announcement and the docs say different things.
When Orchestrator reaches Google AI Pro, or which “additional models and plans” it means.
How fast Hydrogen spends Antigravity’s quota, against 3.8 Flash or Argon.
What happens to Gemini 3.8 Flash. It stays on every Antigravity plan and on the API, with no end date announced.

Hydrogen in Enjoy

Enjoy runs Antigravity on your Google account, through the server Google offers apps such as Zed and JetBrains IDEs, beside Claude Code, Codex and the other labs’ agents, each on its own plan. Its model list shows “the Gemini models Google offers your account” and is fetched from the agent each time you open the settings, so if Google offers your account Hydrogen there, it appears without an Enjoy update. Enjoy’s effort levels for Antigravity are Model default, Low, Medium and High, so Hydrogen’s Extra level isn’t one of Enjoy’s choices; Google’s own app and CLI offer it. Google documents choosing Orchestrator in its app and CLI, and doesn’t mention other apps.

When Hydrogen gets stuck, you can switch agents in the middle of the conversation and hand the job to Claude Code or Codex on the plans you already have. Enjoy is a monthly subscription, with the first seven days free (pricing). It doesn’t add Hydrogen to an account Google hasn’t given it to.

We’ll update this memo when Google settles who gets Hydrogen in Antigravity, or brings Orchestrator to more plans.

9.Questions.

Is Gemini 4 Hydrogen free?: In most places, yes. Free Gemini app accounts get it “via auto model selection when it’s the appropriate model for the prompt,” Search uses it for everyone, and the Gemini API’s free tier includes it, rate-limited and with your content “used to improve our products.” Google AI Plus, at $4.99 a month, lists “Access to Gemini 4 Hydrogen” in the app.
Can I use Gemini 4 Hydrogen in Antigravity for free?: Google’s pages disagree. Its announcement and Antigravity’s blog say Hydrogen is “rolling out across paid Google AI Pro and Ultra subscribers,” but Antigravity’s models page marks it for Free and Google AI Plus too, and its plans page lists it among the “baseline features” of “All plans.” Check your model picker; Google hasn’t explained the difference.
How much does Gemini 4 Hydrogen cost?: On the Gemini API’s paid tier, $1.50 per million input tokens and $7.50 per million output tokens, thinking included, flat across its 1 million-token window, and $0.075 for cached input. That’s twice Gemini 3.8 Flash’s price through December 31, 2026, the same from January 1, and below Argon’s at every date.
Is Gemini 4 Hydrogen better than Gemini 4 Argon?: On Artificial Analysis’s Intelligence Index, slightly: 51 to Argon’s 50 at each model’s high setting, and 54 at Hydrogen’s xhigh. Its tasks cost $1.87 there against Argon’s $9.95, and took 6 minutes against 30. Google still puts Argon and Carbon ahead for “deep, long-horizon reasoning on your hardest problems.”
What is Orchestrator in Antigravity?: A choice in Antigravity’s model picker that pairs two models in one conversation, with Gemini 4 Carbon planning and reviewing and Hydrogen doing the edits, builds and tests. Google says it reached “near-Carbon task success rates” with 24% less quota and 19% faster. Antigravity’s models page lists it for Google AI Ultra only, and you can’t set its thinking level.
Should I switch from Gemini 3.8 Flash to Hydrogen on the API?: Google recommends it, calling Hydrogen the “Recommended upgrade from Gemini 3.8 Flash and earlier Gemini 3 and 2.5 models.” Budget for twice 3.8 Flash’s token price until January 1, 2027, and check the migration list first. Hydrogen defaults to `high` thinking, rejects `minimal`, ignores `temperature` and doesn’t support explicit caching.