Skip to content

Fast & private Try it How it works Pricing Library Setup guide Get it for Mac

Home
Library
Facts
Conformer encoders with transformer decoders have the best average WER

Fact · Speech recognition

The Open ASR Leaderboard paper observes that Conformer-based encoders paired with transformer-based decoders achieve the best average word error rate.

Checked against its source on October 16, 2026

In the source's own words

“Conformer-based encoders paired with transformer-based decoders achieve the best average WER”

Open ASR Leaderboard: Towards Reproducible and Transparent Speech Recognition Evaluation, Hugging Face and collaborators, arXiv, 2025, Abstract.

Open the original

What it measures: From 86 systems on 12 datasets. The same paper says CTC and TDT decoders are faster.
What it does not show: The ranking changes as models are added; check the live leaderboard.

Source

Open ASR Leaderboard: Towards Reproducible and Transparent Speech Recognition EvaluationPreprint · Hugging Face and collaborators, arXiv, 2025

Cited on

Speech Recognition: What It Is and How Well It WorksTopic
How Speech Recognition Works, From Sound to TextGuide
Speech Recognition Models: How Neural ASR WorksGuide
Speech Recognition Benchmarks and LeaderboardsGuide
Which neural network is best for speech recognition?Question
ConformerGlossary

About Pricing Library Questions Glossary Support Privacy Policy Terms of Service Get in touch

Yap is a product of O-mega Enterprise Inc. © 2026

Made for thoughts that move faster than your fingers.

Conformer encoders with transformer decoders have the best average WER | Yap