According to Simon Willison, a well-known developer and blogger, he's updated his OpenAI WebRTC audio playground with support for GPT-Realtime-2, OpenAI's latest voice model. The model brings what OpenAI describes as GPT-5-class reasoning capabilities to voice interactions. The updated tool lets you paste document text and have audio conversations about that content directly in your browser. Willison originally built the playground in December 2024 to try out OpenAI's realtime audio API, and has been waiting for GPT-Realtime-2 to arrive in the ChatGPT mobile app—which still hasn't happened.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton