Models

Each agent can access multiple AI models from different providers. You control which models are available and which one is used by default.

​

View Available Models

Open your agent → Models. You’ll see models grouped by provider — only providers you’ve connected in Secrets will appear here. Each provider section shows:

• Models currently enabled for this agent
• Which model is set as the default
• An + Add Model button to enable more

​

Add a Model

1. Click + Add Model in the provider section
2. Select from available models
3. The model is now available for this agent

Your agent can switch between any enabled models during conversations.

​

Set the Default Model

The default model is what your agent uses unless you specify otherwise. One model will show a DEFAULT badge. To change it, click the model you want as default and select Set as Default.

​

Remove a Model

Click the × next to any model to remove it from this agent. You can always add it back later.

​

Switch Models in Chat

In the chat interface, use the model dropdown in the top right to switch between enabled models mid-conversation. This is useful for:

• Using a faster model for simple tasks
• Switching to a more capable model for complex reasoning
• Testing how different models handle the same prompt

​

Supported Models

Available models depend on your connected providers: Anthropic — Claude Opus, Sonnet, Haiku (various versions) OpenAI — GPT models (including via Codex subscription) OpenRouter — Access to models from multiple providers Gemini — Google’s Gemini models Venice — Privacy-focused models