DialBGM (One-Line Summary) - Kendall Tau: leaderboard
Agreement with human music rankings on DialBGM (2026) when the model sees a one-sentence GPT-4o summary instead of the full dialogue: its ordering of four candidate background-music clips per dialogue is compared with the human ranking over 1,200 dialogues. Kendall tau-b from -1 to 1; higher is better.
Metric: Kendall tau-b (from -1 to 1) between the model's ranking of the four clips and the human ranking, averaged over dialogues, with a one-sentence GPT-4o summary of the dialogue as input, on DialBGM, 1,200 everyday multi-turn dialogues each paired with four candidate music clips ranked by human annotators for suitability as background music, the model scoring each audio clip; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 0.1808 | #276 |
| 2 | Gemini 2.5 Flash | 0.1765 | #463 |
| 3 | Qwen2.5-Omni | 0.1674 | #1588 |
| 4 | GPT-Audio | 0.1625 | #607 |
| 5 | Audio Flamingo | 0.1004 | |
| 6 | Music Flamingo | 0.0965 | #866 |
| 7 | Phi-4 Multimodal Instruct | 0.0241 | #4154 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=dialbgm-one-line-summary-kendall-tau · How It Works · Data refreshed daily, snapshot 2026-11-21.