Models

It helps to know that under the hood, your agents are powered by advanced AI language models. So every Agent ultimately calls an LLM. We won’t dive into technical model names (no need to worry about GPTs or anything like that). Instead, our platform keeps it simple with three AI levels: Express, Core, and Ultra. Each level corresponds to a different power of the AI brain:

Express — A lightweight, fast model. Great for quick, simple answers and basic tasks.
Core — A balanced model. Good all‑around capabilities for most customer interactions.
Ultra — The most powerful model. Thinks step by step and excels at understanding complex questions and giving detailed answers, though it may be a bit slower.

No matter which level you choose for your agent, the way to get great results is the same: provide clear instructions. Learn more about how to write great instructions

Tip: Select the right model tier for your use case is important. You can start with Express, and if the Agent is not able to handle the task, you can switch to Core or Ultra. This way you can ensure that the MAS run as quick as possible. We suggest you to use Ultra if the task is critical.

Model Settings

We provide a few settings to configure the model of an Agent. You can set them in Advanced Settings of an Agent.

Field	Type	Notes
`temperature`	`number`	Creativity vs. determinism.
`topP`	`number`	Nucleus sampling.
`frequencyPenalty`	`number`	Penalise repeated tokens.
`presencePenalty`	`number`	Encourage new tokens.
`maxTokens`	`number`	Maximum tokens in the response.