AudioBench - YouTube SDS: leaderboard

AudioBench YouTube SDS slice measuring spoken-dialogue summarization over YouTube audio.

Metric: Mean Score (%). Source: github.com. Saturation forecast: Estimated already saturated. 8 models tracked.

Top models

#ModelScoreOverall rank
1Gemini 1.5 Flash65.97#618
2Qwen2-Audio-7B-Instruct51.58#880
3Qwen-Audio-Chat43.88#1244
4SALMONN-7B31.28#1293

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=audiobench-youtube-sds · How It Works · Data refreshed daily, snapshot 2026-10-20.