OpenAI adds GPT-Live-Transcribe and GPT-Transcribe for lower-latency audio transcription
OpenAIDevs · x · 2026-07-29
OpenAI introduced two new transcription models in its API:
- GPT-Live-Transcribe for low-latency live transcription
- GPT-Transcribe for asynchronous transcription of completed audio files and batch workloads
The company says both models better use context and improve accuracy on real-world audio across accents and languages. On its new Context Aware ASR benchmark, GPT-Live-Transcribe’s semantic accuracy rose from 38.5% without free-form context to 44.6% with it.
Other reported results:
- On Common Voice across 22 languages, GPT-Live-Transcribe reached 19.70% WER vs 20.33% for GPT-Realtime-Whisper-1
- On Real-World Audio Recording across 9 languages, it posted 9.60% WER vs 11.65% for GPT-Realtime-Whisper-1
OpenAI also says builders can improve live transcription by supplying free-form recording context, keywords, expected input languages, and earlier transcribed turns.
More from Models
- Weird Model Behavior: Opus 5 Loves Saying 'Sabotage Test' — emax · 2026-07-30
- Engineer Debunks Kimi K3 Memory Claims: Small State ≠ Flash Offload — AccBalanced · 2026-07-30
- Dev Critiques Claude Opus: Brilliant but Lacks Rigor, Only Does What It Wants — heyneighbor · 2026-07-30
- OpenAI: GPT-5.6 Fuses Frontier Intelligence with Efficiency — Outside-Iron-8242 · 2026-07-30
- OpenAI Says GPT-5.6 Sol Self-Optimizes: 20% Lower Serving Costs — OpenAI · 2026-07-30
- Opus 4.8 Emits 6x More Tokens Per Turn for Denser Deliberation — jyangballin · 2026-07-30