RESCUE-BENCH - Intervention Time Prediction: leaderboard
Relation-aware multi-party emotional support: predicting when a supporter should intervene in the conversation. F1, higher is better. Source: arxiv.org/abs/2609.09657.
Metric: F1 (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 Plus | 94.6 | #136 |
| 2 | DeepSeek V3.2 | 86.86 | #211 |
| 3 | Kimi K2.5 | 85.98 | #153 |
| 4 | GPT-4o | 84.6 | #357 |
| 5 | DeepSeek V4 Flash | 84.22 | #155 |
| 6 | DeepSeek V4 Pro | 83.28 | #112 |
| 7 | MiniMax-M2.5 | 82.25 | #316 |
| 8 | DeepSeek R1 | 80.36 | #265 |
| 9 | Qwen 3 Max | 75.14 | #213 |
| 10 | Qwen Plus | 68.87 | #393 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rescue-bench-intervention-time-prediction · How It Works · Data refreshed daily, snapshot 2026-11-06.