Fact · Speech recognition
3.29percent
On the authors' combined nine-accent English test set, Whisper-large-v3 fully fine-tuned on the training data had an average word error rate of 3.29 percent.
Quote found word for word on the live source, last checked October 19, 2026
In the source's own words
“Fully fine-tuned Whisper 1541.62 1541.62 8.29 3.29 3.31 1.93 4.46 3.72 4.77 3.40 1.43 3.60 11.28”
Mixture of LoRA Experts with Multi-Modal and Multi-Granularity LLM Generative Error Correction for Accented Speech Recognition, Bingshen Mu et al., arXiv, 2025, Table II.
Open the original- What it measures
- Table II of the preprint. The test set is the authors' combined multi-accent English set of nine accents (AU, CA, CN, DE, IN, JP, UK, US, ZA). The AVG column is the average word error rate over those sets.
- What it does not show
- Authors' own test set and runs. Only the systems in this table are comparable with each other.
In the data
A row of Speech Recognition Accuracy: Word Error Rates, as Whisper fully fine-tuned: 3.29 percent.
- System
- Whisper-large-v3, fully fine-tuned
- Kind
- Research system
- Test set
- Combined multi-accent English test sets (9 accents)
- Language
- English
- Speech
- Accented or dialect
- Year
- 2025