Fact · Speech recognition
The Open ASR Leaderboard reports two metrics: word error rate for transcription quality and inverse real-time factor for inference speed.
Checked against its source on October 17, 2026
In the source's own words
“We report results on two metrics: word error rate (WER) for comparing transcription quality, and inverse real-time factor (RTFx) for comparing inference speed.”
Open ASR Leaderboard: Towards Reproducible and Transparent Multilingual and Long-Form Speech Recognition Evaluation, Hugging Face and collaborators, arXiv, 2025, Section 2.3 Metrics.
Open the original- What it measures
- The leaderboard reports one accuracy number and one speed number for each system, on the same audio.
- What it does not show
- Accuracy and speed are measured separately. A fast system is not necessarily an accurate one, and the leaderboard reports both so the trade-off is visible.