Model Providers¶
What it's for¶
Model Providers is meant to be the place to compare the AI services behind your agents. These are the language model (LLM), text-to-speech (TTS), and speech-to-text (STT) providers. The page has four tabs: LLM Providers, TTS Providers, STT Providers, and Benchmarks.
This page is a preview
Right now the provider cards, costs, latency figures, and benchmark charts on this page are fixed sample content. They do not reflect your workspace or your agents' real usage. The Add Provider, Configure, and Primary & Fallback Configuration controls do not save anything. Changes you make here have no effect on your calls.
Where to see the models your agents actually use¶
Each agent's real model and voice settings come from Retell AI. You can see them on the agent's page.
- Go to Voice Agents and open an agent with View Details.
- The summary bar at the top shows the agent's LLM and Voice.
- Open the LLM Settings tab to see the Provider, Model, and any Fallback LLM.
- Open the Voice & Speech tab to see the Voice Provider and Voice ID.
- Open the Transcription tab to see the speech-recognition model.
To change which model or voice an agent uses, edit the agent in Retell, then click Sync from Retell on the Voice Agents page.
See real cost and usage¶
For figures based on your actual calls, use these instead of the Benchmarks tab:
- The Cost Tracker dashlet on the Dashboard shows usage cost over time.
- Each agent's Call Settings tab shows Cost per Minute, Total Tokens Used, and Total Cost Incurred.