Gemini 4 Hydrogen
Google released Gemini 4 Hydrogen on November 12, 2026, generally available in the Gemini API and rolling out to the Gemini app and Google Search for all users.
- Lab
- Google DeepMind
- Release date
- Status
- Generally available
- API model ID
- gemini-4-hydrogen
- Context window
- 1,048,576 tokens
- Max output
- 65,536 tokens
The Gemini API changelog lists Gemini 4 Hydrogen (gemini-4-hydrogen) as generally available on November 12, 2026, built for speed and efficiency in everyday coding and agentic workflows, available on the free tier and the paid tier. It adds a fourth xhigh thinking level alongside low, medium and the default high, and is the default model for the Antigravity agent in Gemini Managed Agents. The model page lists text, image, video, audio and PDF input with text output, a unified 1,048,576-token context window and a default output limit of 65,536 tokens (up to 1,048,576 with long decode). The model card gives a knowledge cutoff of March 2026 and says Hydrogen is adapted from Gemini 4 Carbon.
Google’s announcement says Hydrogen is rolling out globally: the Gemini API in AI Studio and Android Studio, Google Antigravity for paid Google AI Pro and Ultra subscribers, Gemini for business, Gemini in Google Sheets, the Gemini app model menu and Google Search for all users. See also the Gemini models list.
- Modalities
- Text, image, video, audio and PDF input; text output
- Knowledge cutoff
- March 2026
- Availability
- Gemini API free and paid tiers (AI Studio, Android Studio); Gemini for business; Gemini app and Google Search for all users; Google Antigravity for paid Google AI Pro and Ultra subscribers; Gemini in Google Sheets. Max output 65,536 tokens by default, up to 1,048,576 with long decode.
- Official source
- ai.google.dev
- Last checked
Questions
When was Gemini 4 Hydrogen released?
What is the Gemini 4 Hydrogen model ID?
gemini-4-hydrogen.Is Gemini 4 Hydrogen on the free tier?
What is the Gemini 4 Hydrogen knowledge cutoff?
Source
Release dates, model IDs and limits come from each lab’s own announcement, changelog, documentation or deprecations page, linked on every model with the date it was last checked. How we pick and check sources: editorial standards.
