AudioBench - IMDA 30s Dialogue Summarization: leaderboard
AudioBench IMDA 30s Dialogue Summarization slice measuring summarization quality for spoken dialogue transcripts and audio-derived conversations.
Metric: Mean Score (%). Source: github.com. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | cascade_whisper_large_v2_gemma2_9b_cpt_sea_lionv3_instruct | 54.52 | |
| 2 | cascade_whisper_large_v3_llama_3_8b_instruct | 53.13 | |
| 3 | Qwen2-Audio-7B-Instruct | 38.91 | #880 |
| 4 | Qwen-Audio-Chat | 31.3 | #1244 |
| 5 | SALMONN_7B | 18.34 |
Interactive version: theaggregate.ai/benchmark?slug=audiobench-imda-30s-dialogue-summarization · How It Works · Data refreshed daily, snapshot 2026-10-20.