Gemini 4 Argon reaches developers: hard code and long workflows
Google made Gemini 4 Argon generally available through the API. It’s aimed at code migrations, agents, math and science.
On October 13, Google released Gemini 4 Argon to developers. The Gemini API changelog calls it the company’s most capable model so far.
What’s new
Google positions the model for long, multi-step tasks that need deep reasoning and precise structure:
- code migrations and complex coding;
- agentic workflows;
- math and scientific research;
- enterprise knowledge work and cybersecurity defense.
It also arrived in GitHub Copilot the same day.
How to try it
- Developers: the paid tier of the Gemini API in Google AI Studio and Android Studio, with the model ID
gemini-4-argon. - Google AI Ultra subscribers: AI Studio and Google Antigravity, the agent-first development platform.
Limitations
The announcement has no pricing, benchmarks or context window size: those live in a separate developer guide. The free API tier isn’t listed.
Prepared with AI from public sources.