Atom
English

Models & Reasoning

Click the model name beside the attachment button to choose your AI provider, model, and reasoning level without leaving the chat.

Choose a provider and model

  • Use the provider icons to switch between Claude Code, Codex, and OpenCode.
  • Choose a model from the list below the provider icons.
  • OpenCode includes search and context-window details for larger model catalogs.

Atom remembers your model and reasoning choices for each provider.

Set the reasoning level

Drag the reasoning slider to control how much work the model puts into a response. Lower settings are faster for straightforward changes; higher settings are better for difficult or ambiguous work but can take longer and use more of your provider allowance.

Available levels depend on the selected model and can include Low, Medium, High, xHigh, and Max.

Codex fast mode

The lightning button appears for supported Codex models. Fast mode runs them at 1.5× speed and consumes credits at 2.5× the standard rate. Atom remembers your choice between chats.

Fast mode uses ChatGPT credits and included limits faster. When Codex is authenticated with an API key, OpenAI’s API pricing applies instead.