Updated daily at 05:00 UTC
Benchmark
Artificial Analysis Speech to Text (AA-WER)
Measures: Speech-to-text APIs and models transcribe three held-out sets (50% AA-AgentTalk, a proprietary set of agent-style speech; 25% cleaned VoxPopuli; 25% cleaned Earnings22), heavily normalized before scoring; ranked by word error rate, alongside speed factor and price.
- Maintainer
- Artificial Analysis
- Introduced
- —
- Status
- Active
- Saturation
- Near-ceiling
- Typical provenance
- Independent
- In scoring set
- Yes
Leaderboard
Every result we hold, whoever ran it. Only independent and third-party rows count towards a score.
No results to show.