# Select a model and runtime environment

The model handles reasoning and generation. The runtime environment executes code, commands, and file-processing work.

## Select an agent

Before sending the first message, select the agent name at the lower left of the input box. Choose an assistant from the menu using its name and description. Alternatively, open **Agents** and select **Use** on the target assistant.

Agent selection may be unavailable in an existing conversation. Start a new conversation from the desired assistant when you need to switch.

## Select a model

Select the model name at the lower right of the input area to open the list of models available in the workspace. Choose a model before sending your request. Availability varies by workspace and deployment.

For complex reasoning, coding, or multi-file analysis, compare available models on the same small task and check accuracy and elapsed time. Automation has its own execution-model field; check it separately when creating a task.

## Select a runtime environment

Open the runtime-environment selector in the input area. Keep **Auto** when you do not need a specific environment, or choose a connected, available environment from the list.

If an environment is missing, open [Resource Center → Runtime Environments](../runtime-environments/index.md) and check whether it exists, is connected, and is accessible to the current workspace. For network failures, inspect that environment's network settings. Changing the model does not configure network access.