Gemini 4 Argon now available on AI Gateway
By Steven Van ·
The model accepts text, images, PDFs, and video in a 1M token context window, with three thinking levels for coding and reasoning tasks.
Developers can use Google’s Gemini 4 Argon for coding and multi-step agent tasks through Vercel AI Gateway. The model supports tool calling and accepts text, images, PDFs, and video within a 1M token context window.
Coding and document analysis
The release note describes Gemini 4 Argon as able to build features, fix bugs, and research unfamiliar codebases across longer tasks. It also reads dense diagrams, spreadsheets, layout-heavy documents, and video, including financial analysis spread across several spreadsheet tabs.
Choose a thinking level
Gemini 4 Argon uses low, medium, and high thinking levels instead of thinking budgets. Set the level with reasoning in the AI SDK, reasoning_effort in Chat Completions, or reasoning.effort through the Responses API.
Connect an app or coding agent
Use google/gemini-4-argon as the model name through the AI SDK, Chat Completions, Responses, or Anthropic Messages APIs. For Responses requests, point the OpenAI SDK to https://ai-gateway.vercel.sh/v1 and authenticate with an AI Gateway API key or Vercel OIDC token.
To use the model in Claude Code, Codex, Cursor, or another connected coding agent, install the latest Vercel CLI, run vercel ai-gateway setup, then select google/gemini-4-argon in the agent.