88Z AI, 8 May — OpenAI has released three real-time voice models that enhance reasoning and tool use, offer low-cost multilingual translation, and provide low-latency transcription, all delivered via the Realtime API with end-to-end emotion preservation.
At a glance
- GPT-Realtime-2 features GPT-5-level reasoning and tool use
- GPT-Realtime-Translate supports over 70 languages with minimal cost
The update
OpenAI has officially launched three new real-time voice models designed to enhance the depth and efficiency of voice interactions. GPT-Realtime-2 integrates GPT-5-level reasoning capabilities and tool usage, significantly improving complex task handling. GPT-Realtime-Translate focuses on bridging language gaps, supporting real-time translation across more than seventy languages at a cost of approximately 0.25 yuan per minute, which is a hundredfold reduction compared to human simultaneous interpretation. Additionally, GPT-Realtime-Whisper is dedicated to providing low-latency speech transcription. All three models are accessible via the Realtime API and utilize an end-to-end processing architecture that effectively preserves speaker tone and emotional nuance, offering users a more natural and fluid voice experience.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- openAISource published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.