OpenAI launches three audio models for real-time voice tasks to enhance conversational software agents
OpenAI introduced three audio models for its developer platform, enhancing voice software capabilities.The new models aim to create more conversational agents capable of real-time task completion during live interactions.
New audio models
The models include GPT-Realtime-2, GPT-Realtime-Translate and GPT-Realtime-Whisper, available for testing in the developer playground.
GPT-Realtime-2 handles complex requests and maintains context in long voice sessions.
GPT-Realtime-Translate facilitates translation from over 70 languages into 13, suitable for customer support and educational settings.
GPT-Realtime-Whisper provides live speech-to-text capabilities, generating captions and meeting notes during discussions.
Customers testing these models comprise online real estate marketplace Zillow, travel agency Priceline, and telecommunications firm Deutsche Telekom.
Pricing for GPT-Realtime-2 starts at $32 per million audio input tokens, while GPT-Realtime-Translate is $0.034 per minute and GPT-Realtime-Whisper is $0.017 per minute.
Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents.
Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold.
Now available in the API… pic.twitter.com/2DY1LU2vO8
— OpenAI (@OpenAI) May 7, 2026
Pricing details
OpenAI’s latest advancements could significantly impact various industries, enhancing efficiency and user interactions.
The focus on real-time capabilities reflects the growing demand for integrated voice technology solutions.
Developers are encouraged to explore these tools to improve their voice-based applications.
The integration of these models could represent a significant step towards more intuitive software agents.
