Skip to content

Using Models

Multi uses a profile-based system to manage your AI providers and models. Each profile stores a provider, model, and configuration - and you can switch between profiles instantly.

A profile is a saved configuration that includes:

  • Provider - Which AI service to use (e.g., Claude Code, Anthropic, OpenRouter)
  • Model - Which specific model to use (e.g., claude-sonnet-4-5, gemini-2.5-pro)
  • Settings - Provider-specific options (API key, temperature, max tokens, etc.)
  1. Open the Multi panel → click Settings (gear icon)
  2. Click Add Profile
  3. Give it a name (e.g., “Claude Sonnet”, “Gemini Free”, “GPT-4”)
  4. Select a provider
  5. Configure the provider settings (API key, model, etc.)
  6. Save

Creating a new profile in settings

Click the profile selector or type > at the bottom of the Multi panel, which shows the currently selected profile name, to switch between your saved profiles. Your default profile is used for new tasks.

Set a default profile in settings - this is the profile used when you start a new task without explicitly selecting one.

  • Create multiple profiles - One for each use case (e.g., “Fast edits” with Haiku, “Deep work” with Opus)
  • Try different providers - The same task can produce very different results across models
  • Use free providers for exploration - Start with CLI providers for quick iteration
  • Match model to task - Don’t use Opus for a simple rename; don’t use Haiku for a complex refactor

You can monitor token usage per task directly in the Multi panel to keep track of costs and consumption.

Viewing token usage for a task