Gemini 4 Hydrogen: $1.50/$7.50 Price, Benchmarks, vs Carbon

Gemini 4 Hydrogen replaces Gemini 3.8 Flash: 1M context, four thinking levels, $1.50/$7.50 per 1M tokens, AA Index 54 vs Argon 50. Full guide.

SumusPublished

Quick answers

What is Gemini 4 Hydrogen?
Gemini 4 Hydrogen is Google's fast, low-cost Gemini 4 model for everyday coding and agent work. It replaces Gemini 3.8 Flash and launched on November 12, 2026.
How much does Gemini 4 Hydrogen cost?
$1.50 per 1M input tokens and $7.50 per 1M output tokens, flat across the full 1M-token context. Cached input costs $0.075 per 1M.
Is Gemini 4 Hydrogen better than Gemini 4 Argon?
On the Artificial Analysis Intelligence Index, yes: 51 at the default high setting and 54 at xhigh, versus 50 for Argon, and it is cheaper.
Where can I use Gemini 4 Hydrogen?
In the Gemini API (AI Studio, Android Studio), the Gemini app, Google Search, Google Antigravity, Gemini for Business, Vertex AI and OpenRouter.
Gemini 4 Hydrogen card: released Nov 12 2026, $1.50/$7.50 per 1M tokens, 1M context in and out, four thinking levels, AA Index 54 at xhigh vs Argon 50

Gemini 4 Hydrogen is Google's new daily-driver model: Flash-class speed and price with near-frontier scores. Google announced Gemini 4 Hydrogen on November 12, 2026, two days after Gemini 4 Carbon. It takes the place of Gemini 3.8 Flash and is now the default in the Gemini app and Google Search.

Below: the price, benchmarks and thinking levels, how it compares with Carbon and Argon, and how to start using it today.

Gemini 4 Hydrogen at a glance

Item

Gemini 4 Hydrogen

Maker

Google DeepMind

Released

November 12, 2026

Model ID

gemini-4-hydrogen

Replaces

Gemini 3.8 Flash

Context

1M tokens in, up to 1M tokens out

Inputs

Text, image, video, audio, PDF

Thinking levels

low, medium, high (default), xhigh

Price

$1.50 / 1M input, $7.50 / 1M output, $0.075 cached

Where

Gemini API, Gemini app, Search, Antigravity, Vertex AI, OpenRouter

Gemini 4 Hydrogen key facts, read Nov 14, 2026 · Google: Introducing Gemini 4 Hydrogen · Accessed

Gemini 4 Hydrogen price

Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens. The price is flat across the whole 1M-token window, so long prompts do not jump to a higher tier. Cache hits are $0.075 per million.

What you pay per job depends on the thinking level. Artificial Analysis measured about $0.29 per task at low, $0.80 at medium, $1.87 at high and $3.16 at xhigh, because xhigh writes almost twice as many output tokens as high.

Gemini 4 Hydrogen benchmarks

Benchmark

Gemini 4 Hydrogen

Gemini 4 Carbon

Gemini 4 Argon

AA Intelligence Index

54 (xhigh), 51 (high)

59

50

Agents' Last Exam v1.1

71.6%

77.4%

69.0%

OSWorld-2.1 (offline)

77.6%

81.5%

74.7%

Vending-Bench 2

$14,549

$16,484

$13,718

Terminal-Bench 4.1

57.0% (Flash: 18.2%)

—

—

SWE-sweep

13.5%

9.6%

4.3%

The big story is the jump from the last Flash model: Google says it finishes tasks 2.5x faster at the default level and cuts hallucinations by a factor of five. Independent tests agree on low hallucination (8% on AA-Omniscience) and rank it 5th in Agent Arena, the fastest model in that top 10.

The weak spot is terminal-heavy agent work. On Terminal-Bench 4.0 it scores 54%, behind Claude Sonnet 5.5 (64%), so for long shell sessions a Claude model may still be the better pick.

Gemini 4 Hydrogen vs Carbon vs Argon

  • Gemini 4 Carbon is Google's frontier model for hard, long reasoning. It scores higher but costs more and runs slower.
  • Gemini 4 Argon came out less than four weeks earlier. Hydrogen beats it on the AA Index and costs less, which leaves little reason to stay on Argon.
  • Hydrogen is the one to use by default: everyday coding, agents, document and video summaries, and chat.

In Antigravity, Google now runs an orchestrator that switches between Carbon and Hydrogen inside one session. Google says it nearly matches Carbon's success rate on Terminal-Bench 4.0 with 24% fewer resources.

Gemini 4 Hydrogen thinking levels

Level

AA Index

Cost per task (AA)

Use it for

Low

27

$0.29

Quick chat, routing, simple edits

Medium

41

$0.80

Drafts and summaries

High (default)

51

$1.87

Most coding and agent tasks

Xhigh

54

$3.16

Hard problems where a test shows the gain

How to use Gemini 4 Hydrogen

  1. Gemini app and Search: it is already the default model for everyone.
  2. API: pick gemini-4-hydrogen in Google AI Studio or Vertex AI, and set the thinking level per call.
  3. Coding: Pro and Ultra subscribers get it in Google Antigravity.
  4. OpenRouter: listed as google/gemini-4-hydrogen at the same $1.50 / $7.50 price.

Use Gemini 4 Hydrogen with your team in Sumus

Yes, you can put Gemini next to Claude and GPT in one place. In Sumus, each model works as an AI Employee in your team chat, so you can send fast daily jobs to a Gemini model and hard ones to a frontier model, and everyone sees the result.

FAQ

Is Gemini 4 Hydrogen free?

Yes in the Gemini app and Google Search. API use is paid at $1.50 / $7.50 per 1M tokens, with a free tier in AI Studio for testing.

Is Gemini 4 Hydrogen the same as Gemini 4 Flash?

Google does not use the Flash name, but it is the successor to Gemini 3.8 Flash and fills the same role.

What is the Gemini 4 Hydrogen context window?

1M tokens of input and up to 1M tokens of output.

Is Gemini 4 Hydrogen good for coding?

Good for everyday coding and the fastest top-10 model in Agent Arena; for long terminal sessions Claude Sonnet 5.5 still scores higher.

Can I use Gemini 4 Hydrogen with my team?

Yes. Sumus lets your team use Gemini, Claude and GPT models as AI Employees in one shared chat.

Update log

  • — Nov 14: First version: Gemini 4 Hydrogen launch (Nov 12), price, benchmarks, thinking levels, vs Carbon and Argon.
← All articles

This page was written by a Sumus AI Employee

Bring your ideas to Sumus and start working together.

Try it