Fact · Accessibility and health
51percent word error rate
In the same 2026 study, every commercial system exceeded 51 percent word error rate on severe dysarthria.
Checked against its source on October 16, 2026
In the source's own words
“severe dysarthria exceeded 51% WER for every system”
Benchmarking Commercial Speech Recognition and Multimodal Large Language Models on Dysarthric Speech, Ali Alsayegh and Tariq Masood, International Journal of Intelligent Systems, 2026, Abstract.
Open the original (DOI 10.1155/int/6065038)- What it measures
- Severe dysarthria in the study means 4 speakers from the TORGO corpus. The authors also tested prompts; the prompt results are separate from this default-setting figure.
- What it does not show
- Four speakers, one corpus and default settings. Prompting changed the figures for some models, so this is not the best each system can do.
Source
Cited on
- Speech-to-Text APIs Compared: Price, Accuracy, SpeedGuide
- Word Error Rate (WER): How to Calculate and Read ItGuide
- Speech Recognition Benchmarks and LeaderboardsGuide
- Speech Recognition for Stuttering and DysarthriaGuide
- Can voice recognition understand dysarthric speech?Question
- How accurate is voice recognition software?Question
- What is a good word error rate?Question
- What is the most accurate speech-to-text model?Question
- Which speech recognition apps work for atypical speech?Question
- Which speech-to-text API is the best?Question