Finally, ChatGPT Voice Will Stop Interrupting You Mid-Sentence
OpenAI says the new model will interrupt less, add real-time translation and visuals, and route harder queries to stronger text models.
- On Wednesday, OpenAI unveiled GPT-Live-1, its "smartest voice model" yet, for ChatGPT Voice across iOS, Android, and the web. GPT-Live-mini-1 serves free users while the main model powers paid tiers.
- Replacing a turn-based architecture that frequently struggled with natural flow, the new full-duplex framework allows simultaneous listening and speaking. OpenAI product lead Atty Eleti explained that the architecture processes incoming and outgoing streams of data simultaneously.
- The model incorporates conversational fillers like "mhmm" and "yeah" to indicate active listening while translating speech in real-time. Users can now instruct the assistant to remain silent until explicitly called upon, a feature unavailable in previous versions.
- For complex reasoning or web searches, the system automatically routes queries to GPT-5.5, enabling rapid transitions from research to verbal presentation. Built-in guardrails provide "expert-vetted crisis helpline support" and steer interactions away from harmful content during high-risk situations.
- Positioned as a companion rather than a traditional chatbot, OpenAI aims to boost mainstream appeal and advance toward AGI. By incorporating humanlike intricacies like breath intakes and laughter, the company suggests that "true AGI" depends on humanlike voice abilities.
31 Articles
31 Articles
Got ChatGPT’s new voice mode? Here's how to check — and 5 things you should try first
GPT-Live is rolling out to all ChatGPT users nowIt can both talk and listen at the same time for more natural chatsReal-time translations are also now possibleChatGPT has a shiny new AI voice model called GPT-Live, which has a number of helpful tricks — including being able to listen and talk at the same time. It's rolling out to all ChatGPT users now, though OpenAI has acknowledged a number of early bugs.While free users and users on a paid pla…
OpenAI launches new GPT-Live Voice models for more natural AI conversations
OpenAI has introduced its new GPT-Live-1 and GPT-Live-1 mini conversational AI models, designed to make ChatGPT voice interactions more natural, faster and human-like. The new models improved turn-taking, real-time responses, live translation capabilities and better contextual understanding.
OpenAI introduces GPT-Live for ChatGPT Voice: Can the latest update replace real human interactions?
Now you are going to enjoy talking to ChatGPT, as it will speak and listen at the same time, like an individual. Yes, OpenAI has introduced a new voice model, GPT-Live, that can engage in real-life human-AI conversations. The capability is rolling out globally immediately and is now powering ChatGPT Voice. OpenAI has announced the rollout of two new voice models of GPT-Live – GPT-Live-1 and GPT-Live-1 mini – that can listen and talk simultane…
Coverage Details
Bias Distribution
- 54% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
