> ## Content Index
> Fetch the complete content index at: https://www.theleftshift.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI Launches gpt-realtime: Its Most Advanced Speech-to-Speech AI Model Yet
- URL: https://www.theleftshift.com/openai-launches-gpt-realtime-its-most-advanced-speech-to-speech-ai-model-yet/
- Published: 2025-08-29T05:38:35.000Z
- Updated: 2025-08-29T05:38:35.000Z
- Description: To enhance personalisation, OpenAI has introduced two exclusive new voices, Cedar and Marin, available only with gpt-realtime.
- Author: The Left Shift Bureau
- Tags: AI Models, AI Startups

OpenAI has[ unveiled](https://openai.com/index/introducing-gpt-realtime/?ref=theleftshift.com) **gpt-realtime**, its most advanced **speech-to-speech AI model**, promising a leap forward in natural and expressive voice interactions. 

The model, now available through the **Realtime API**, is designed to power next-generation voice agents with unprecedented speed, accuracy, and fluency.

This follows the [release of GPT-5](https://www.theleftshift.com/openai-releases-gpt-5-with-major-upgrades-not-agi-yet/), the most advanced model from OpenAI so far. During the same period, the San Francisco-based startup also [released two open-source models. ](https://www.theleftshift.com/openai-launches-open-source-models-tops-hugging-face-leaderboard/)

Unlike traditional systems that stitch together separate speech-to-text and text-to-speech pipelines, gpt-realtime directly processes and generates audio within a **single integrated model**, reducing latency and preserving nuance in speech. 

This breakthrough allows conversations to feel more fluid and human-like, whether for customer support, real-time translation, or interactive voice assistants.

> The Realtime API is officially out of beta and ready for your production voice agents!  
>  
> We’re also introducing gpt-realtime—our most advanced speech-to-speech model yet—plus new voices and API capabilities:  
>  
> 🔌 Remote MCPs  
> 🖼️ Image input  
> 📞 SIP phone calling  
> ♻️ Reusable prompts [pic.twitter.com/fX5yvt0CDD](https://t.co/fX5yvt0CDD?ref=theleftshift.com)
> 
> — OpenAI Developers (@OpenAIDevs) [August 28, 2025](https://twitter.com/OpenAIDevs/status/1961124915719053589?ref%5Fsrc=twsrc%5Etfw&ref=theleftshift.com)

The model also demonstrates significant improvements in following complex instructions, interpreting system prompts, and **switching seamlessly between languages mid-sentence**. 

It can handle precise tasks such as reading legal disclaimers verbatim or repeating alphanumeric strings—key features for enterprise-grade applications.

To enhance personalisation, [OpenAI](https://www.theleftshift.com/openai-claims-major-drop-in-hallucinations-with-gpt-5/) has introduced two exclusive new voices, **Cedar and Marin**, available only with gpt-realtime.

“Voice AI is moving beyond novelty to production-ready systems,” OpenAI said in its announcement. “gpt-realtime brings the naturalness, reliability, and low latency needed to deploy at scale.”