By model
| Model | /min | Total |
|---|---|---|
| GPT-Transcribe | $0.0045 | $0.27 |
| GPT-Live-Transcribe | $0.0170 | $1.02 |
If you'll summarize the text with another model afterwards, add that cost: an hour of speech is about 11,000 tokens.
Transcribing 1 hour of audio with GPT-Transcribe ($0.0045 per minute) costs $0.27. In real time, with GPT-Live-Transcribe, it would be $1.02.
| Model | /min | Total |
|---|---|---|
| GPT-Transcribe | $0.0045 | $0.27 |
| GPT-Live-Transcribe | $0.0170 | $1.02 |
If you'll summarize the text with another model afterwards, add that cost: an hour of speech is about 11,000 tokens.
Yes, OpenAI's transcription models handle dozens of languages.
Estimate based on public prices as of October 2026. Token counts vary with the text and the model; check the official pricing page before you decide.