OpenAI globally launched GPT-Live on Tuesday, integrating direct search capabilities and visual answers into its conversational artificial intelligence for ChatGPT Voice users.
The upgrade allows ChatGPT Voice to process queries with advanced reasoning models and web search, providing more dynamic and informative user interactions, according to OpenAI.
GPT-Live-1 will serve as the default for Go, Plus, and Pro subscribers, while a lighter version, GPT-Live-1 mini, will be available to free users. These models operate in full-duplex, enabling simultaneous audio input and speech output to minimize interruptions during conversations.
Users can now receive visual cards within ChatGPT Voice for various topics, including real-time weather updates, stock market data and sports scores. OpenAI stated that this feature enhances the utility of voice interactions by providing immediate visual context.
The new system also allows users to adjust reasoning levels, offering choices between “Instant,” “Medium,” and “High.” These settings correspond to different processing intensities, such as GPT-5.5 Instant or GPT-5.5 Thinking, depending on the complexity of the query.
OpenAI reported that more than 150 million people currently utilize ChatGPT’s voice and dictation features weekly. The company confirmed it is actively developing support for voice with video and screen sharing, although these functionalities are not available at launch.
Source: Search Engine Journal
