The model is your agent’s brain. It’s what understands what callers say, figures out how to respond, and generates the words your agent speaks. Different models have different strengths — some are faster, some are more expressive, some are more affordable.
Location: Prompt Section (top bar) → Model dropdown
Available Models
| Model | Provider | Type | Credit Usage |
|---|
| Electron | Smallest AI | Traditional | Lower |
| GPT-4o | OpenAI | Traditional | Moderate |
| GPT-4.1 | OpenAI | Traditional | Moderate |
| GPT Realtime | OpenAI | Emotive | Higher |
| GPT Realtime Mini | OpenAI | Emotive | Higher |
Emotive models have emotional awareness — they pick up on caller tone and respond with natural expression.
Getting started? GPT-4o is a great all-rounder — reliable, capable, and well-tested.
Try Electron. It’s our own model, built specifically for voice. Fast, affordable, and optimized for real-time conversations.
You can switch models anytime — nothing breaks when you do.