Voice chat models are designed to enable voice-based interactions with AI systems. Unlike traditional text-only assistants, these models can generate natural-sounding speech as responses, creating a more immersive and human-like conversational experience. Some models accept text input and respond with voice, while others can process both speech and text, allowing users to talk directly to the model or type messages depending on the use case.

Depending on the model, you may have access to settings for bitrate, output audio formats (often including lossless options), stream vs. non-stream modes, as well as a variety of voices and ways to customize or modify them.

Several models that were originally listed in our Text Models (LLM) section should also be included in this category:

• gpt-4o-audio-preview
• gpt-4o-mini-audio-preview