OpenAI has rolled out a major update to its developer platform by adding a set of advanced voice-focused tools aimed at improving conversational applications. The new capabilities allow apps to handle spoken interactions, convert speech into text, and translate conversations instantly.
Among the additions is GPT-Realtime-2, an upgraded voice model built to deliver more natural dialogue and improved reasoning during live interactions. The company says the system can better manage complex user requests and maintain smoother conversations compared to earlier versions.
OpenAI also introduced a real-time translation feature capable of processing dozens of spoken languages while delivering translated responses with minimal delay. Alongside it, the new GPT-Realtime-Whisper model enables instant transcription during active conversations, giving developers access to live speech-to-text functionality.
The company believes the technology could support industries such as customer support, digital media, online learning, and event services. OpenAI also stated that safety protections were added to reduce risks related to spam, scams, and abusive content, while pricing for the tools depends on either usage time or token consumption.