AIToday
Large Language ModelsITmedia AI+Published: Nov 16, 2026, 19:01 JST

Google opens Gemini 4 Hydrogen to all users

Send on LINE
Google opens Gemini 4 Hydrogen to all users

3 Key Points

  1. What happened

    On November 12, Google released Gemini 4 Hydrogen, a speed-tuned model built on Gemini 4 Carbon, and made it available to all users in the Gemini app, Google Search, the API and Antigravity.

  2. Why it matters

    The model is weighing becoming the default for free-tier Gemini app users when Google's auto-selection deems it a fit, and paying Google AI subscribers can pick it explicitly.

WHO IT HITSDevelopers building on the Gemini API and Google AI Studio get a low-cost model for coding, debugging and multi-step research tasks, while Google Workspace and Gemini for business administrators can offer it to enterprise users.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Gemini 4 Hydrogen is the third entry in the Gemini 4 series, following Argon in September and Carbon in October. Google positions Carbon as its highest-performance model, and Hydrogen is tuned from it for speed and efficiency rather than deep reasoning, which the company leaves to Carbon and Argon. That division of labor extends to Google's developer tool Antigravity, where a new Orchestrator feature automatically switches between a top model and Hydrogen within a single session.

The model's intended roles run from focused coding and debugging to multi-step research and knowledge work such as turning PDFs, spreadsheets, audio and video into sourced reports. In a demo, Hydrogen generated a Google Slides deck from a YouTube lecture video and a design template, then checked and fixed overlapping text on its own. Google also says the model makes drafting emails, building study guides and making product comparison tables in the Gemini app more convenient.

On safety, Google's model card compares Hydrogen with the previous fast model, Gemini 3.8 Flash: safety on inputs including images improved, but refusals of legitimate requests also rose, especially for inputs containing images. Google also cautions that response delays and timeouts occasionally occur, and that higher thinking levels can consume many tokens.

FAQ
How much does Gemini 4 Hydrogen cost through the API?
Input costs $1.5 per 1M tokens ($0.075 for cached input) and output costs $7.5 per 1M tokens. That matches Gemini 3.8 Flash's default rate, though 3.8 Flash is half price through December 31.
Can free users access Gemini 4 Hydrogen?
Yes. Free Gemini app users get it through the automatic model-selection feature when Google judges the prompt a good fit, and the free API tier includes it at no charge.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleWhite House to name Nvidia, Amazon, Microsoft, OpenAI, Anthropic AI 'National Champions'