# Qwen3-ASR model evaluation: Voice Benchmark Profile

> Qwen’s model card reports word error rates for Qwen3-ASR 1.7B and 0.6B across AMI, Earnings22, GigaSpeech, LibriSpeech, SPGISpeech, and VoxPopuli.

- Scope: Mean WER over seven English test sets plus per-set values
- Measurement lanes: Voice experience
- Primary metric: Word error rate (lower is better)
- Available evidence: Exact figures published on the official model card
- Owner: Qwen
- Source snapshots refreshed: 2026-09-11

## Interpretation limit

Provider-run evaluation on public English test sets.

## Primary sources

- [Owner page](https://huggingface.co/Qwen/Qwen3-ASR-1.7B-hf)

Canonical page: https://benchlm.ai/voice-benchmarks/qwen3-asr-1-7b-launch-evaluation