Aircaps launches A5S Async V1, claims #1 on VoiceArena with 41% fewer errors
massimosgrelli · x · 2026-11-10
Aircaps introduces A5S Async V1, billed as the world's most accurate speech recognition model for speaker-attributed transcription. It ranks #1 on VoiceArena ahead of ElevenLabs, AssemblyAI, Deepgram, NVIDIA and Meta, with the vendor claiming the next-best model makes 41% more errors across 5 public benchmarks. Acknowledging every lab claims #1, the team open-sourced its benchmark for verification.
Related event: AirCaps A5S Async v1 Tops VoiceArena Speech Recognition Leaderboard(3 posts)→