Whisper Large v3 vs AssemblyAI Universal
Two speech-to-text models compared on price, context, capabilities and measured speed. External prices are public list rates verified 2026-09-16; Inference APIs prices are live.
| Whisper Large v3 | AssemblyAI Universal | |
|---|---|---|
| Provider | OpenAI | AssemblyAI |
| Model id | openai/whisper-large-v3 | universal |
| Price | $0.002 / audio minute | $0.0025 / audio minute |
| Capabilities | Speech to text, 99 languages, Timestamps, SRT / VTT | Speech to text, Diarization, Speech understanding |
| Weights | Open | Closed |
| Availability | Available on Inference APIs | Available |
| Rate limits | No daily caps; pay per request | Listed as $0.15 per hour (≈ $0.0025 / minute). |
| Real-time factor | 0.047× (19.5s clip in 0.92s) | not measured |
Cost for 10 hours of audio
Whisper Large v3 is about 20% cheaper for this workload at list price. Price is only part of it: free tiers and usage tiers cap how much you can send per day, and a retired model is unavailable at any price. The "Rate limits" row above is the practical difference for a production app.
Measured speed
Speed figures for Inference APIs models are medians of three runs from a European client on 2026-09-16, against the public endpoint, using a ~120-word generation prompt (chat), a 300-character paragraph (speech) or a 19.5-second clip (transcription). External models are not measured here; treat "not measured" as unknown, not slow.
When to pick which
- Whisper Large v3 — OpenAI's multilingual speech recognition model on the standard /v1/audio/transcriptions endpoint — same request shape as OpenAI and Groq.
- AssemblyAI Universal — Listed as $0.15 per hour (≈ $0.0025 / minute).
Try Whisper Large v3
Same OpenAI request shape; the change is the base URL and key. Full parameters, aliases and pricing on the model page.
External prices and limits come from the linked provider pages and were verified on 2026-09-16. Spotted a change? Tell us.