# GPT-Realtime-Translate GPT-Realtime-Translate is [[OpenAI]]'s live translation model, released on 7 May 2026 alongside [[GPT-Realtime-2]] and [[GPT-Realtime-Whisper]]. You stream speech in; it streams translated speech out, plus the transcript, while the speaker is still talking. It understands more than 70 input languages and speaks 13 output languages. The use case is any conversation where two people don't share a language: support calls, sales calls, tutoring, events, or a product video translated as it plays (Vimeo's launch demo). Deutsche Telekom was testing it for multilingual customer support at launch. ## Specifications and pricing | | GPT-Realtime-Translate | |---|---| | Model ID | `gpt-realtime-translate` | | Endpoint | `v1/realtime/translations` (its own session type, not a voice-agent turn loop) | | Input / output | Audio in; audio and text out | | Context window | 16,000 tokens | | Max output | 2,000 tokens | | Languages | 70+ input, 13 output | | Price | $0.034 per minute of audio | | Rate limits | 50 (tier 1) to 850 (tier 5) minutes of audio per minute | Billing by the minute makes it easy to estimate: one hour of translated audio costs about $2.04. ## How you use it You set the target language on the session. One session translates into ONE language, so the docs say to create one session per target language. For a two-person call, that means two sessions: caller A into B's language, and B into A's. In group rooms, the number of sessions is roughly the number of active speakers times the number of target languages, which adds up quickly. ## Claims and caveats BolnaAI, an Indian voice AI company, says it got 12.5% lower word error rates than any other model it tested in Hindi, Tamil and Telugu, with lower fallback rates. That's a vendor-quoted customer claim; OpenAI didn't publish a translation benchmark in the launch post. OpenAI's own guide lists what to test before going live: language-pair quality, code-switching and mixed-language conversations, among others. It's closed and API-only, like the rest of OpenAI's realtime lineup. ## References - OpenAI, "Advancing voice intelligence with new models in the API": https://openai.com/index/advancing-voice-intelligence-with-new-models-in-the-api/ - Model docs: https://developers.openai.com/api/docs/models/gpt-realtime-translate - Realtime translation guide: https://developers.openai.com/api/docs/guides/realtime-translation - Pricing: https://developers.openai.com/api/docs/pricing - TechCrunch, "OpenAI launches new voice intelligence features in its API": https://techcrunch.com/2026/05/07/openai-launches-new-voice-intelligence-features-in-its-api/ ## Related - [[GPT-Realtime-2]] - [[GPT-Realtime-Whisper]] - [[OpenAI Realtime API]] - [[2026-05-07 GPT-Realtime-2 - OpenAI's voice model gets GPT-5-class reasoning]] - [[Speech-to-Text (STT)]] - [[Text-to-Speech (TTS)]]