OpenAI has launched new voice AI features for its API platform, helping developers build applications that interact with users, transcribe speech, and translate languages. The new GPT-Realtime-2 model offers realistic voice simulation, enabling natural conversations with users. Unlike the previous version, this model possesses GPT-5 level reasoning capabilities and is designed to process more complex requests. This is reported by Techcrunch.com reports .
Additionally, the company introduced the GPT-Realtime-Translate feature. It provides real-time translation services during conversations and supports over 70 input and 13 output languages. Furthermore, the GPT-Realtime-Whisper tool offers live speech-to-text transcription, recording interactions instantly.
OpenAI representatives state that these new models transform voice interfaces from simple Q&A systems into tools capable of performing complex tasks—listening, analyzing, and acting. These technologies are expected to revolutionize sectors such as customer service, education, media, and content creation.
Regarding security, the company has implemented special protection systems to prevent abuse, fraud, and spam. If harmful content rules are violated during a conversation, the system automatically terminates the interaction. The new voice models are included in the OpenAI Realtime API, with pricing based on usage time or token consumption.
Read “Zamin” on Telegram!
