AI
OpenAI launches GPT-Live, a full-duplex voice model rolling out in ChatGPT today
OpenAI launched GPT-Live, a new generation of voice models it describes as its smartest yet, beginning a rollout today across iOS, Android, and web. Built on a full-duplex architecture that lets the model listen and speak simultaneously, GPT-Live is designed for more natural back-and-forth conversation, real-time translation, and a more defined personality than prior voice models. For complex tasks requiring web search or deeper reasoning, it can silently delegate to a frontier model in the background and weave the result back into the conversation.
Sam Altman called the experience "magical and real," adding that he has always preferred typing to talking to an AI but expects that preference to shift. OpenAI Developers separately confirmed that GPT-Live-1 and a smaller GPT-Live-1 mini variant are coming to the API soon, with a sign-up form open for developers and enterprises wanting early access.
Cursor partners with xAI to train Grok 4.5, its first model built beyond software engineering
Cursor announced a partnership with xAI to train Grok 4.5, calling it their most powerful model to date and the first they have built for use cases beyond software engineering. The model is available now inside Cursor with double usage credits for the first week. The company clarified that Grok 4.5 and its existing Composer 2.5 model represent different weight classes and will both continue to be offered going forward.
This briefing is part of the archive.
Get the app for full archive access — every daily briefing, sourced and attributed.