# OpenAI launches three real-time audio models for voice assistants and translation tools

> OpenAI released three real-time audio models - GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper - built for voice assistants and customer support. All three are available now via API.

- Type: AI news
- Page: https://completeaitraining.com/news/openai-launches-three-real-time-audio-models-for-voice/
- Published: 2026-05-08
- Updated: 2026-05-08
- Author: Jeroen Erne
- Jobs: Customer support
- Reading time: 1 min

### OpenAI Releases Real-Time Audio Models for Customer Support

OpenAI has launched three new real-time audio models designed for voice assistants and customer support systems. The models-GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper-are now available through the company's API.

GPT-Realtime-2 brings GPT-5-level reasoning to voice conversations, allowing systems to handle complex requests and maintain natural dialogue without delays. The model understands context and can reason through multi-step tasks as users speak.

#### What This Means for Support Teams

For customer support operations, the update addresses a core challenge: handling inquiries faster while maintaining conversation quality. Voice interactions let customers explain problems naturally without typing, which matters when they're multitasking or need help in real time.

GPT-Realtime-Translate handles multilingual support without separate processing steps. GPT-Realtime-Whisper provides live transcription, converting speech to text as conversations happen.

Support teams can use these models to reduce response times and cover more languages with fewer resources. The real-time processing means customers don't wait for systems to process their words before responding.

#### Building Voice Support Systems

Developers can integrate these models into existing support platforms or build new voice-first systems. The API approach means organizations choose which models fit their workflow-some may need translation, others live transcription, others reasoning-heavy problem solving.

Learn more about [AI for Customer Support](https://completeaitraining.com/tag/customer-support/) and [Speech-To-Text](https://completeaitraining.com/tag/speech-to-text/) capabilities for your team.

---

This is the Markdown form of https://completeaitraining.com/news/openai-launches-three-real-time-audio-models-for-voice/. The whole site is mapped at https://completeaitraining.com/llms.txt.
