Choose your agent’s LLM

The LLM is the model that does your agent’s thinking and writes its answers. Your choice affects how it responds and how much usage a conversation takes.

Hosted agents

Choose from all the leading frontier models, plus a range of leading open-weight models.

Open your agent’s Manage page and choose Brain in the sidebar.

Choose a model from the Model dropdown. Leave it on Default to follow Omnimatch’s default model as it changes, or choose a model by name to keep that selection. The change takes effect on your next message.

The small bars compare how much allowance the models use for the same amount of text. Longer conversations and reasoning can use more. See Account usage for what the bars mean.

Fall back to house models lets another model answer if the selected model is overloaded or not responding. The model that answers determines the usage charged.

On-premise agents

On-premise uses the model configured for your installation, shared across its agents. Your administrator chooses it during setup and can change it under System administration → LLM.

You can connect a model running on your own infrastructure, an external provider or Omnimatch models where your licence includes access. The installation’s model takes precedence over individual agent model selections. See Choosing a model for on-premise for setup details.