AI library

Source · docs

Gemini 4 Argon model page (Gemini API docs)

Author
Google
Published
13 October 2026
Added
15 October 2026
Original
https://ai.google.dev/gemini-api/docs/models/gemini-4-argon

Summary

The Gemini API documentation page for gemini-4-argon, checked on 2026-10-15. It labels the model Stable, and the Gemini API changelog entry dated 2026-10-13 says Argon "reached general availability". The page lists its limits and supported features; the publication date is taken from the changelog entry, as the page itself says only "October 2026".

Key points

  • Status: Stable (not a preview). GA on the Gemini API on 2026-10-13, per the changelog.
  • Context: 1,048,576 tokens, shared between input and output.
  • Output: 65,536 tokens by default for synchronous requests; up to 1,048,576 with "long decode".
  • Inputs: text, image, video, audio, PDF. Output: text only. No Live API, image, or audio generation.
  • Thinking levels: low, medium, high.
  • Tools: function calling, structured output, code execution, file search, URL context, Search and Maps grounding; computer use is still a preview feature.
  • Caching: implicit caching only; explicit (context) caching is not supported.
  • Batch API, Flex, and Priority inference are supported.
  • No knowledge cutoff or price is shown on this page.

Why it matters

It settles whether Argon is a preview: it is a stable, GA model on the Gemini API. The lack of explicit caching matters for cost planning with large, repeated contexts.