Fact · Speech recognition
The Open ASR Leaderboard reports the inverse real-time factor (RTFx) rather than the real-time factor.
Checked against its source on October 17, 2026
In the source's own words
“We report the inverse real-time factor, rather than real-time factor”
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 compares 86 open-source and proprietary systems on English and multilingual tests, reporting WER for quality and RTFx for speed.
- What it does not show
- The paper reports RTFx for its own hardware and batching setup. Values are comparable within the leaderboard, not across unrelated tests.