RESCUE-BENCH - Relation Pattern Prediction: leaderboard
Metric: Accuracy (%). Source: arxiv.org. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Pro | 45.6 |
| 2 | Kimi K2.5 | 45.26 |
| 3 | Qwen 3 Max | 43.22 |
| 4 | Qwen 3.5 Plus | 40.27 |
| 5 | MiniMax-M2.5 | 39.85 |
| 6 | DeepSeek V4 Flash | 39.39 |
| 7 | GPT-4o | 38.69 |
| 8 | Qwen Plus | 38.27 |
| 9 | DeepSeek R1 | 37.89 |
| 10 | DeepSeek V3.2 | 36.08 |
Interactive version: theaggregate.ai/benchmark?slug=rescue-bench-relation-pattern-prediction · How It Works · Data refreshed daily, snapshot 2026-09-21.