How to Configure Different (Cheap) LLM for OpenClaw’s Heartbeat

08 Feb, 2026

By using faster + cheaper model for OpenClaw heartbeat, we can save the performance and also LLM cost.

Yeah, you don’t need your powerful LLM for that use case.

You can configure different LLM for heartbeat by adding the configuration at openclaw.json file.

{
	"agents": {
		"defaults": {
			"heartbeat": {
				"model": "google/gemini-3-flash",
				"every": "30m"
			}
		}
	}
}

Reference