Qwen2-Audio-7B-Instruct — benchmark results

Provider: Alibaba. Released 2024-07-31. Access: Open.

Unified ELO 1304 ± 26, rank #1522 of 1776 rated models, from 72 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AudioBench - IMDA GR Sentence67.36Mean Score (%)100
AudioBench - MuChoMusic71.61Score (%)100
AudioBench Music65.89Average Score (%)100
AudioBench - SEAME Mandarin WER55.23WER (%) (lower is better)90
AudioBench - AISHELL ASR ZH WER9.26WER (%) (lower is better)88.9
AudioBench - IMDA Part 2 ASR WER19.06WER (%) (lower is better)88.9
AudioBench - IEMOCAP Gender95.57Mean Score (%)87.5
AudioBench - IMDA GR Dialogue61.28Mean Score (%)87.5
AudioBench - MELD Sentiment55.81Mean Score (%)87.5
AudioBench - WavCaps Captioning30Mean Score (%)87.5
AudioBench - WavCaps QA46.34Mean Score (%)87.5
AudioBench - VoxCeleb Gender99.39Mean Score (%)85.7

Interactive version: theaggregate.ai/model?slug=qwen2-audio-7b-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.