Try PixVerse C1 Reference-to-Video Model from PixVerse →
Sign InStart Building

Task History

Click to see output list

No tasks yet

Go to Models
Realtime Conversation models

Realtime voice conversation models.

Low-latency voice models that listen and answer in real time, for voice agents, live translation and phone assistants.

Showing 9 of 9 models
gemini 3.8 live - AI model cover image

gemini 3.8 live

google

Google Gemini 3.8 Live is a low-latency audio model for voice agents. It streams natural 24 kHz speech, supports tool calling, and optional Search grounding.

Realtime Conversation
Run
gpt-live-1 - AI model cover image

gpt-live-1

openai

GPT‑Live‑1 is OpenAI’s full‑duplex voice model for real-time conversation. It streams speech and transcripts and can delegate reasoning and tools to a backend model.

Realtime Conversation
Run
gpt-realtime-translate - AI model cover image

gpt-realtime-translate

openai

Stream live speech and get translated speech plus live captions with low delay. Made by OpenAI for calls, meetings, broadcasts, and video chat translation.

Realtime Conversation
Run
gpt-realtime-2.1-mini - AI model cover image

gpt-realtime-2.1-mini

openai

Build low-latency voice agents with OpenAI’s GPT‑Realtime 2.1 Mini. It takes live audio or text and replies with spoken audio plus a transcript. ([developers.openai.com](https://developers.openai.com/api/docs/models/gpt-realtime-2.1-mini))

Realtime Conversation
Run
gpt-realtime-2.1 - AI model cover image

gpt-realtime-2.1

openai

OpenAI's GPT Realtime 2.1 runs speech-to-speech conversations with configurable reasoning and tool calls. It’s built for voice agents that handle noise and interruptions.

Realtime Conversation
Run
Realtime Conversational AI - AI model cover image

Realtime Conversational AI

elevenlabs

A real-time voice conversation tool using ElevenLabs' AI voice agents. Customize voices, behaviors, and languages for interactive AI experiences.

Realtime Conversation
Run
gpt-realtime-mini - AI model cover image

gpt-realtime-mini

openai

GPT Mini Realtime enables low-latency, bidirectional streaming for voice and text. Build interactive, responsive AI experiences that feel natural and immediate.

Realtime Conversation
Run
gpt-realtime - AI model cover image

gpt-realtime

openai

GPT Realtime enables low-latency, bidirectional streaming for voice and text. Build interactive, responsive AI experiences that feel natural and immediate.

Realtime Conversation
Run
PersonaPlex-Realtime - AI model cover image

PersonaPlex-Realtime

nvidia

Convert speech to speech with customizable voices using PersonaPlex. Supports various audio formats and offers control over text temperature and audio top K settings.

Realtime Conversationspeech-to-textaudio
Run

How to pick a realtime voice model

Realtime models hold a spoken conversation: they listen, answer with speech and let the user interrupt. They power voice agents, phone assistants and live translation.

Voice agents with tools

GPT Realtime 2.1 supports tool calls and adjustable reasoning, and is built to cope with noise and interruptions. Gemini 3.8 Live streams 24 kHz speech, calls tools and can ground answers with Search.

Listen and talk at once

GPT-Live-1 is full duplex, so it can listen while it talks, and it can hand reasoning and tool work to a backend model.

Frequently asked questions

How do I connect to a realtime model?

Start the session with a POST to the model’s Run endpoint, open a WebSocket with the token you get back, then stream audio in both directions. The realtime voice guide walks through it.

How is a conversation billed?

Per turn, taken from your credits while the session runs, so you can stop whenever you like.

What audio format should I send?

The realtime guide’s example sends 24 kHz PCM audio to an OpenAI model. Each model lists the formats and sample rates it accepts as parameters on its page.