OpenAI Whisper

OpenAI offers two STT paths. GPT-4o Transcribe (API) posts 8.9% WER on Artificial Analysis benchmarks — the highest accuracy of any managed API. gpt-4o-mini-transcribe at $0.003/min is the best price-performance option for OpenAI ecosystem teams. gpt-realtime-whisper handles live transcription at $0.017/min. Self-hosted Whisper Large-v3 (MIT license, open weights) runs free on your own GPU — the industry baseline open-source model available via HuggingFace, or hosted through Groq, Fireworks, and Replicate at $0.0003–$0.003/min. Note: Whisper Large-v3 is no longer the accuracy leader on any major benchmark, though it remains the dominant self-hosted choice by usage volume. Best for: batch transcription where accuracy matters most, feeding transcripts into GPT workflows, and self-hosted deployments requiring full control.