---
title: Dutch ASR Leaderboard
sdk: gradio
sdk_version: "6.13.0"
app_file: app.py
pinned: false
license: apache-2.0
tags:
  - automatic-speech-recognition
  - hf-asr-leaderboard
  - asr
  - speech
  - speech-recognition
  - leaderboard
  - evaluation
  - benchmark
  - Dutch
  - leaderboard
models:
  - openai/whisper-large-v3
  - CohereLabs/cohere-transcribe-03-2026
  - Qwen/Qwen3-ASR-1.7B
  - nvidia/parakeet-tdt-0.6b-v3
language:
  - nl
metrics:
  - wer
  - rtf
---

# Dutch ASR Leaderboard

Community leaderboard for Dutch Automatic Speech Recognition models.

## Features

- 📊 **Multiple datasets**: FLEURS, VoxPopuli, MLS
- 🔄 **Dynamic filtering**: Toggle proprietary models, select datasets
- 📈 **Ranking options**: Sort by average or individual dataset WER

## Evaluation Setup

Local models are benchmarked on **H100 HBM2e GPUs**.

## How to Submit

1. Run `eval.py` to generate results JSON
2. Add your `results/<model>.json` file
3. Open a PR

See the [GitHub repo](https://github.com/tvosch/Dutch-ASR-leaderboard) for full documentation.