OpenAI Launches GPT-Live to Make ChatGPT Voice Conversations More Natural
OpenAI launches GPT-Live, introducing faster, more natural voice conversations, visual features and enhanced safety for ChatGPT users across platforms.
SociapaHQ
OpenAI has unveiled a new generation of voice models to make conversations with its chatbot, ChatGPT, more natural, responsive and human-like. And the rollout comes with an upgraded ChatGPT Voice experience, with quicker response times, better listening and integrated visual abilities on iOS, Android and the web.
The rollout includes two models: GPT-Live-1 and GPT-Live-1 mini, and they are rolling out globally. The company said the new voice system is based on a full-duplex architecture that allows ChatGPT to listen and speak at the same time, letting users interrupt naturally and have more fluid, real-time conversations.
GPT-Live enables a smoother conversation compared to traditional voice assistants, which wait for the user to finish speaking before replying, and also handles more complex tasks such as web browsing and advanced reasoning with OpenAI’s latest GPT-5.5 model running in the background.
Also Read: Truecaller Challenges TRAI Over Anti-Spam Rules
Enhanced Voice AI:
Users can select Instant, Medium or High reasoning modes, depending on whether they prioritise speed or more in-depth analysis. In addition to voice responses, ChatGPT can now also show live visual cards for things like weather and stocks and sports scores. It continues to support file uploads and image inputs and memory features, too.
OpenAI said GPT-Live-1 will be the default voice model for Go, Plus and Pro subscribers, while Free users will be provided with GPT-Live-1 mini. The company will also offer the new voice models to developers through its API at a later date.
GPT-Live also has natural verbal acknowledgements like “mhmm” and “got it” to help make interactions more conversational, pauses to let users finish speaking without interrupting and filters background noise for better accuracy. OpenAI has also redesigned the nine built-in voices of ChatGPT to hold more expressive and context-aware conversations.
Also Read: Apple Signs $30 Billion Chip Manufacturing Deal with Broadcom
The company also said it has enhanced safety measures as part of the launch, including checks for risks concerning self-harm, emotional dependency, violence and inappropriate content. The update also includes better protections for younger users through parental controls and age-appropriate safeguards.
To reduce the risk of voice cloning and impersonation, GPT-Live currently only supports the company's pre-selected verified voices.
OpenAI said real-time video conversations and screen-sharing aren’t currently available in GPT-Live, but are in the works as the company continues to expand the platform’s multimodal capabilities.
/afaqs/media/post_attachments/aa8f8388a655c7c57255593c3e76a44e58db6bf8f23b35539e52907cdb3a2a29.png){kind=small}
