Google launched Gemini 4 Hydrogen on November 12, 2026, a fast Gemini 4 model rolling out in Google Search, the Gemini app, the Gemini API and other Google products. In a post on The Keyword, Koray Kavukcuoglu, SVP of Google DeepMind and Chief AI Architect at Google, said Hydrogen is "rolling out globally across our developer, enterprise, and consumer surfaces". Google says: "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."

#What changed?

Google positions Hydrogen against its frontier Gemini 4 models. "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."

Google calls Hydrogen "a major standalone upgrade for teams migrating from Gemini 3.8 Flash". Its Gemini API guide describes the model as a "Recommended upgrade from Gemini 3.8 Flash and earlier Gemini 3 and 2.5 models." Ars Technica's Ryan Whitwam reported: "This model has replaced Gemini 3.8 Flash across the company’s products." Google has not published a shutdown date for 3.8 Flash.

Google describes "Hydrogen’s four thinking levels" as "low, medium, high (default), and a new xhigh tier". The API guide says: "By default, the model uses the high thinking level and a 64K synchronous output token limit", and that "Input, thinking, and output tokens share one 1,048,576-token window."

In Google Antigravity, Google says it is "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." According to Google, "On Terminal-Bench 4.0, Orchestrator reached near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster."

In the Gemini app, Google says Hydrogen "breaks down your goal step-by-step, pulling relevant context from your uploads, Google apps like Gmail and Google Drive, and Connected Apps like Airtable and Dropbox", including through Google's Connected Apps Platform preview.

On Search, a separate Keyword post by Robby Stein, VP of Product, Google Search, says Hydrogen is "available globally in Google Search for all users". Google also says people will see "more images, structured responses, comparison tables, and custom tools built for you." Ars Technica reported that "the responses in both AI Overviews and AI Mode will have more visual elements."

Google says Hydrogen is "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". According to Google, the model scores 51 on the Artificial Analysis Intelligence Index at its default high thinking level, versus 50 for Argon, and 54 at xhigh. Google says "our partners have seen Hydrogen run 2x to 2.8x faster while reducing token usage by 12% to 23%, compared to 3.8 Flash."

On safety, Google says Hydrogen "reduces overreach in agentic tasks by up to 81% and under-service by up to 79%". It also says that "Against indirect prompt injection attacks, Hydrogen is more than 2.3x more resilient on the Gray Swan IPI benchmark". Google says its updated safeguards are more precise at distinguishing "between real biosecurity risks and legitimate medical or scientific questions", "keeping false refusals on HealthBench Pro under 0.4%."

#How much does it cost?

Google's paid-tier API price is $1.50 per million input tokens and $7.50 per million output tokens, flat across the 1 million-token context window. The API guide also lists cached input tokens at $0.075 per 1M tokens via implicit caching, a 95% discount, and says Gemini 4 Hydrogen does not support explicit caching. Ars Technica reported the price is unchanged from Gemini 3.8 Flash.

#Who can use it and when?

Access in the Gemini app depends on plan, and the sources describe free-tier app access and AI Studio availability differently.

Surface

What the sources say

Google Search

Google says Hydrogen is "available globally in Google Search for all users". Ars Technica reported that everyone will "reportedly get access to Hydrogen in search".

Gemini API and Google AI Studio

Google's API guide says the model "is available on both the free and paid tiers in Google AI Studio and the Gemini API". Ars Technica reported that access is "a bit more restricted in the Gemini app and AI Studio".

Gemini app

Google says "free users will get access to Gemini 4 Hydrogen via auto model selection when it’s the appropriate model for the prompt." Ars Technica reported: "The new model is live for AI Pro and Ultra subscribers, but AI Plus and free users are still stuck with the old 3.5 Flash-Lite model."

Google Antigravity

Google says paid Google AI Pro and Ultra subscribers can try it there. Its API guide says the Antigravity agent uses Gemini 4 Hydrogen by default.

Gemini in Google Sheets

Google says this starts with Google AI Pro and Ultra subscribers, and is also available "in Gemini in Google Sheets for eligible business subscriptions."

Google's developers post tells builders to start with the Gemini API in Google AI Studio and Android Studio, or in Gemini for business, and says paid Google AI Pro or Ultra subscribers can try it in Google Antigravity.

#How to get it

Google says Hydrogen is "available globally in Google Search for all users". In the Gemini app, Google points people to the model drop-down.

For developers, Google's API guide says:

  1. Change the model string. "Change your model string from gemini-3.8-flash to gemini-4-hydrogen."
  2. Check thinking levels. "Hydrogen defaults to high rather than medium, adds xhigh, and rejects minimal with an error."
  3. Plan for caching limits. The guide says Gemini 4 Hydrogen does not support explicit caching.

#How does it compare?

Gemini 4 Carbon went to cyber defenders first. Google still positions Carbon and Argon for deep, long-horizon reasoning, and Hydrogen for rapid execution. Ars Technica reported that the Argon and Carbon rollout "began with cybersecurity researchers", while the limits for Hydrogen "are much more relaxed".

The launch also sits alongside Google's Gemini Decisions API and Decision 1 model, introduced on November 11, 2026.

Our analysis: Hydrogen is the Gemini 4 model Google is pushing to the largest set of surfaces, at a flat API price and with a documented upgrade path from 3.8 Flash. The benchmark and safety figures are Google's own claims, not independent tests. For most consumers, the practical question is which Gemini app plan actually serves Hydrogen. Google says free users can get it through automatic model selection when it is the right model for the prompt. Ars Technica reported that AI Plus and free users are still on 3.5 Flash-Lite. Those accounts do not match, and neither source resolves the other.

#What is still unclear?

Google and Ars Technica still disagree on free-tier access in the Gemini app, and on how open Google AI Studio is. No source gives a shutdown or deprecation date for Gemini 3.8 Flash. Google's API guide calls the free tier rate-limited but does not publish the numeric limits.