OpenAI launches GPT-Live voice models for ChatGPT

Investing.com -- OpenAI released GPT-Live on Wednesday, a new generation of voice models designed to make conversations with artificial intelligence more natural. The models now power ChatGPT Voice.

GPT-Live uses a full-duplex architecture that allows it to listen and speak simultaneously. During conversations, the model can acknowledge input with phrases like "mhmm" or "yeah," engage in rapid exchanges, or remain silent when needed.

The company said GPT-Live can delegate tasks requiring web search, deeper reasoning, or complex work to frontier models running in the background. At launch, GPT-Live will use GPT-5.5 for these tasks. OpenAI plans to update the underlying model as new frontier models become available.

OpenAI is rolling out two versions: GPT-Live-1 and GPT-Live-1 mini. Both are being distributed to ChatGPT users globally starting Wednesday. The company said it plans to make the models available through its API soon.

Previous voice systems relied on cascaded models that processed speech in sequence, or turn-based models that required users to finish speaking before responding. GPT-Live processes input continuously while generating output, allowing it to make interaction decisions multiple times per second.

In head-to-head evaluations, GPT-Live-1 and GPT-Live-1 mini were preferred over Advanced Voice Mode in conversations lasting five to ten minutes. The evaluations measured overall preference, turn-taking, interruptions, conversational flow, and how natural interactions felt.

More than 150 million people use ChatGPT Voice and Dictation features weekly, according to OpenAI.

The new ChatGPT Voice experience includes natural conversations, smarter answers with access to frontier models, improved listening capabilities, and visual responses. Users can choose reasoning levels: Instant for fast responses, or Medium and High for more complex thinking.

ChatGPT Voice can now display visual cards for topics including weather, stocks, and sports while users are talking. The feature continues to support search, memory, images, and file uploads.

GPT-Live-1 will become the default model for Go, Plus, and Pro users, while GPT-Live-1 mini will serve Free users. The models are available on iOS, Android, and ChatGPT.com.

At launch, GPT-Live will not support voice with video or screen sharing in ChatGPT. OpenAI said it is working to add these capabilities. Users can still access previous versions of ChatGPT Voice where these features are available.

You May Also Be Interested In

• Meta to invest $13 billion CAD in first Canadian data center
• ELEKTROS Continues Building Momentum with a Disciplined EV Infrastructure Strategy
• Teako Signs Letter of Intent for Strategic Contribution of 90% Interest in Norwegian High-Grade Silver Projects