Fact · Transcription
0.0015US dollars per audio minute
Fireworks AI's audio transcription launch post lists serverless Whisper-v3-large at $0.0015 per audio minute in its async transcription table.
Quote found word for word on the live source, last checked October 19, 2026
In the source's own words
“[Serverless] Whisper-v3-large (beam size 5) $0.0015 11 sec Variable”
20x faster Whisper than OpenAI - Fireworks audio transcribes 1 hour in 4 seconds, Fireworks AI, Fireworks AI, 2024, Async Transcription Measurements table.
Open the original- What it measures
- The table prices each Whisper model per audio minute, with latency per hour of audio. The serverless rows run on Fireworks-configured GPUs, and dedicated endpoints are listed as Custom.
- What it does not show
- The price comes from a post published December 9, 2024, and the page does not date the table. The current Fireworks pricing page does not list audio or Whisper prices, so this may not be the current rate. Read on 2026-10-19.
In the data
A row of Speech-to-Text API Pricing: Cost per Hour of Audio, as Fireworks Whisper v3 large: 0.09 US dollars per hour of audio.
- Provider
- Fireworks AI
- Product
- Whisper v3 large
- Mode
- Batch (recorded audio)
- Tier
- Serverless
- As of
- 2026-10
- Note
- $0.0015 per audio minute is $0.09 per hour.
Source
Used to calculate
- In the speech-to-text and transcription prices dataset, the 22 published figures for “Batch (recorded audio)” have a median of 0.21 US dollars per hour of audio, ranging from 0.04 to 6.
- The speech-to-text and transcription prices dataset holds 34 data points, each a published figure traced to its source.
- The speech-to-text and transcription prices dataset draws on 15 distinct sources.