AirCaps A5S Async v1 Tops VoiceArena Speech Recognition Leaderboard

AirCaps released A5S Async v1, a speaker-attribution transcription model that tops the VoiceArena leaderboard with cpWER of 23.89% on 139 held-out conversations, claiming a 41% error-rate lead over runners-up like ElevenLabs, AssemblyAI and Deepgram, with the benchmark open-sourced.

2026-11-10 ~ 2026-11-10 · 3 related posts

• Aircaps launches A5S Async V1, claims #1 on VoiceArena with 41% fewer errors — massimosgrelli · 2026-11-10
• AirCaps launches A5S Async v1 speech model with 41% fewer errors, open-sources benchmark — dair_ai · 2026-11-10
• AirCaps A5S Async v1 tops Voice Arena diarization+ASR board at 23.89% cpWER — rdesh26 · 2026-11-10