OpenAI Launches Three Real-Time Voice Models Capable of 'Thinking', Translating, and Transcribing While Listening - AI

7 x 24 Track global technological trends

Hot Topic

Day

News Topic

OpenAI Launches Three Real-Time Voice Models Capable of 'Thinking', Translating, and Transcribing While Listening

11 hour ago / Read about 0 minute

Author：小编

On May 8, 2026, OpenAI released three real-time voice models designed for conversational reasoning, real-time translation, and real-time transcription, unlocking a new generation of voice application formats for developers. GPT-Realtime-2 features GPT-5-level reasoning capabilities, supporting complex request processing; GPT-Realtime-Translation supports over 70 input languages and 13 output languages; GPT-Realtime-Whisper enables low-latency speech-to-text conversion.

Previous page：Neuralink: Developing Surgical Robots to Address B...

Next page：Krisp Launches VIVA 2.0, Reconstructing Voice AI I...

Return to List

Hot Reading

2 day ago

Silicon Valley bets $200M on AI data centers floating in the ocean

1 day ago

AMD posts record first-quarter results, driven by skyrocketing data center CPU demand

1 day ago

Your iPhone will soon give you an important decision to make

1 day ago

Google's Gemma 4 AI models get 3x speed boost by predicting future tokens