Choosing the right model

In your Agent configuration, you can select from models made available by your organization through CARTO AI settings.

Model recommendations

CARTO provides the following guidance to help you choose a model:

Google

  • Complex use cases: Use gemini-2.5-pro for stronger reasoning and tool handling.

  • Smaller / speed-prioritized use cases: Use gemini-2.5-flash for faster responses and lower cost.

OpenAI

  • Complex use cases: Use gpt-4o for higher reasoning quality and robust tool use.

  • Smaller / speed-prioritized use cases: Use gpt-4o-mini for fast, cost-efficient interactions.

Your model selection affects the agent’s reasoning depth, latency, cost, and its ability to handle multi-step tool flows.

Last updated

Was this helpful?