Google's Gemini API model catalog lists Gemini 4 Hydrogen, with the model ID gemini-4-hydrogen, and provides information about its context window, output limits, and input modalities.
Suitable tasks: Everyday coding and agent workflows; this is how the model is positioned in the catalog.
Unsuitable tasks: Configuration decisions that require documentation to specify a dedicated reasoning tier, sampling parameters, or version lifecycle status.
Applicable model version: Gemini 4 Hydrogen. The catalog gives the model ID gemini-4-hydrogen and does not provide a dated or version-suffixed model name.
Applicable client, agent, or API: Gemini API. The documentation page provides a “Try in AI Studio” entry point.
Recommended reasoning tier and parameters: Not specified.
Set the model identifier in a Gemini API request to:
gemini-4-hydrogenThe catalog shows that the model supports text, image, audio, video, and PDF inputs, has a 1M-token context window, and lists a maximum output of 64K / 1M tokens. The page does not explain the conditions for each of the two output limits or provide examples of other request parameters.
Use the model ID gemini-4-hydrogen in a Gemini API request.
Design the request according to the modalities and token limits listed in the catalog.
Before sending a request, confirm which models are available to the project and what account permissions apply; this page does not describe account, regional, or quota restrictions.
The model catalog's “Gemini 4” section explicitly lists “Gemini 4 Hydrogen.”
Model code: gemini-4-hydrogen.
Google describes it as intended for everyday coding and agent workflows, and mentions a shared 1M-token context window and long decoding of up to 1M tokens.
The model card lists Context: 1M tokens; Max output: 64K / 1M tokens; Modalities: Text, Image, Audio, Video, PDF.
The page includes Gemini 4 Hydrogen in the “All Gemini 4 models” table and provides a “Try in AI Studio” entry on its card. This shows that the official catalog publicly lists the model and its model ID; the page does not explicitly specify a publication date, stable or preview status, or other release stage.
This information comes from the official model catalog as it appeared on the collection date. The catalog does not provide a specific Hydrogen version suffix, reasoning tier, sampling parameters, rate limits, or a complete API call example; do not infer these configurations from the available information. The page's general version-naming guidance does not associate a stable, preview, or other status with Hydrogen.
On the collection date, the source URL was opened in a browser, and the page ultimately displayed the “Models” page in Google AI Studio's Gemini API documentation. The Gemini 4 model card and the “All Gemini 4 models” table were checked, and the visible model ID and specifications were recorded.
Gemini 4 Hydrogen