All tasks / Models
Universal-3.6 Pro Streaming
AssemblyAI's real-time speech-to-text model line (Universal-3.6 Pro, WebSocket streaming), billed per session hour.
- AssemblyAI
- Closed
- Released 29 Sep 2026
Where it ranks
- ATranscriptionModels · Cloud · 3.6% on Artificial Analysis Speech to Text: AA-WER Streaming (secondary benchmark)
Facts
- Vendor
- AssemblyAI
- Family
- Universal Streaming
- Released
- 29 Sep 2026
- Weights
- Closed: API and apps only
- License
- proprietary
- API price
- $0.0075 per minute of audio (Official list price $0.45 per hour of streaming session for Universal-3.6 Pro Realtime, divided by 60.)
- Last verified
- 19 Oct 2026
Price from assemblyai.com, 19 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Artificial Analysis Speech to Text: AA-WER Streaming | 3.6% | Universal-3.6 Pro Realtime - Min LatencyAA-WER Streaming index 0.035714 converted to percent; the Max Accuracy variant scores 3.85; AA release date 2026-09-29; page shows no last-updated date. | 19 Oct 2026 | artificialanalysis.ai |
| Coval Benchmarks: Speech-to-Text WER (24h window) | 2.4% | assemblyai / universal-3.6-proCoval STT benchmark, mean WER over the trailing 24h window (snapshot generation 2611, as_of 2026-10-19T10:00Z), n=912 clips; rank 1 of 31; Coval drives AssemblyAI over its WebSocket streaming endpoint. | 19 Oct 2026 | benchmarks-api-6wxgp27p2a-ue.a.run.app |
Where you can use it
No tracked tool lists this model yet.
Earlier versions
- Universal-3.5 Pro Streaming · —