OpenAI vs Fish Audio pricing
STT: OpenAI from $0.003 vs Fish Audio from $0.006.
Last updated · prices checked daily
Speech-to-text: OpenAI vs Fish Audio
OpenAI has the lower listed STT price: $0.003 per minute (gpt-4o-mini-transcribe (estimated)) vs $0.006 for Fish Audio (transcribe-1), about 50% less.
| Vendor | Product / tier | $/min |
|---|---|---|
| OpenAI Visit site ↗ | gpt-4o-mini-transcribe (estimated) Token-billed; OpenAI's own per-minute estimate. | $0.003 |
| OpenAI Visit site ↗ | gpt-transcribe (batch) | $0.0045 |
| Fish Audio Visit site ↗ | transcribe-1 | $0.006 |
| OpenAI Visit site ↗ | gpt-live-transcribe (streaming) | $0.017 |
What it costs at volume (lowest listed rate)
| Minutes | OpenAI | Fish Audio |
|---|---|---|
| 1,000 min | $3.00 | $6.00 |
| 10,000 min | $30.00 | $60.00 |
| 100,000 min | $300.00 | $600.00 |
Cost-per-call calculator
Estimate what one call really costs. Prices come from today's table; LLM and telephony presets are per-minute rates published by Retell and Vapi. Everything is editable.
Cartesia's pricing page puts one minute of speech at 750–800 characters; Rime says about 1,000. The talk share is your own guess at how much of each call the agent is speaking.
FAQ
Is OpenAI or Fish Audio cheaper for speech-to-text?
OpenAI has the lower listed STT price: $0.003 per minute (gpt-4o-mini-transcribe (estimated)) vs $0.006 for Fish Audio (transcribe-1), about 50% less. Prices checked Oct 28, 2026.
How often are these prices updated?
Every day. A script reads each vendor's public pricing page, extracts the listed rates and keeps a dated copy, so price changes show up in the history.