VoiceArena Diarization Bench - cpWER: leaderboard
A transcription system writes a speaker-attributed transcript of each of 139 recorded English conversations (2 to 8 speakers, 115 in person and 24 online, 272,690 reference words), and the number is concatenated minimum-permutation word error rate in percent. Unlike plain word error rate boards such as Open ASR, a word attributed to the wrong speaker counts as an error.
Metric: cpWER on 139 conversations (%). Source: voicearena.com. Saturation forecast: Rough model projection: around 2026. 12 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | AirCaps A5S Async v1 | 23.89 | |
| 2 | Nemotron 3.5 Transcribe | 32.43 | |
| 3 | MOSS-Transcribe-Diarize | 34.73 | |
| 4 | Scribe v2 | 35.73 | |
| 5 | Muse Voice Transcribe | 36.53 | |
| 6 | AssemblyAI Universal-3.6 Pro | 47.81 | |
| 7 | Soniox v5 Async | 48.28 | |
| 8 | Smallest AI Pulse | 50.85 | |
| 9 | Voxtral Small Transcribe 3 | 59.65 | |
| 10 | Voxtral Mini Transcribe 3 | 65.04 | |
| 11 | Speechmatics Enhanced | 66.07 | |
| 12 | Deepgram Nova 3 | 71.89 |
Interactive version: theaggregate.ai/benchmark?slug=voicearena-diarization-bench-cpwer · How It Works · Data refreshed daily, snapshot 2026-11-16.