OpenAI on Thursday introduced three audio models for its developer platform. The goal of the model is to enable voice-based software agents to be more interactive and complete tasks in real-time.
The launch comes as competition in AI-powered voice technology intensifies, with major technology companies investing heavily in multimodal systems that combine text, audio and visual understanding.
The rollout is aimed at developers building conversational AI tools, virtual assistants, customer service platforms and other voice-enabled applications.
Additionally, the launch of the Application Programming Interface API takes ChatGPT-Producer beyond transcription and chat to agents that can listen, translate, and act during live conversations.
OpenAI’s new audio models: GPT-realtime-2, GPT-realtime-translate, and GPT-realtime-whisper
The new models are gpt-realtime-2, gpt-realtime-translate and gpt-realtime-whisper. OpenAI said they are available for testing in its developer playground.
GPT-Realtime-2 is designed to handle complex requests, handle call tools, interruptions, and maintain context in long voice sessions.
The second model supports translation from over 70 languages ​​into 13 output languages, targeting customer support, education and other settings.
GPT-Realtime-Whisper provides live speech-to-text, allowing captions, meeting notes, and workflow updates to be generated while the speaker is talking.
Customers testing the model include online real estate marketplace Zillow, online travel agency Priceline and European telecommunications firm Deutsche Telekom.
Pricing for gpt-realtime-2 starts at $32 per million audio input tokens, gpt-realtime-translate at $0.034 per minute, and gpt-realtime-whisper at $0.017 per minute.
The company said the model is designed to deliver more natural speech, better accuracy and lower latency during live interactions.
