Fact · Speech recognition
7.15percent average WER
On the Open ASR Leaderboard's English short-form tests, NVIDIA Canary 1B v2 had an average word error rate of 7.15 percent, with an RTFx of 749.
Checked against its source on October 17, 2026
In the source's own words
“NVIDIA Canary 1B v2 Yes 7.15 749 FastConformer”
Open ASR Leaderboard: Towards Reproducible and Transparent Multilingual and Long-Form Speech Recognition Evaluation, Hugging Face and collaborators, arXiv, 2025, Table 3, short-form English.
Open the original- What it measures
- The row sits in the same table as the Parakeet and Whisper rows, measured under the same normalization.
- What it does not show
- Average over the leaderboard's English datasets, so it describes English read and conversational audio, not all 25 languages.