MiGUE-Bench - Intra-Document Temporal Relations: leaderboard
A within-document relation task of MiGUE-Bench (2026), a benchmark of event analysis over news documents: accuracy in percent on 300 multiple-choice questions about temporal relations between events of one document, with LLM-built distractors. Higher is better.
Metric: Accuracy (%) on 300 single-document temporal relation multiple-choice questions with LLM-built distractors over news documents; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 | 71 | #156 |
| 2 | Qwen 3 Max | 69.67 | #403 |
| 3 | GLM-4.7 | 69.33 | #366 |
| 4 | Gemini 3 Pro | 68.67 | #137 |
| 5 | Claude Haiku 4.5 | 67.67 | #568 |
| 6 | DeepSeek V3.2 | 66 | #406 |
| 7 | GPT-5.2 Pro | 66 | #124 |
| 8 | Kimi K2 | 59 | #439 |
| 9 | Qwen 3 235B A22B | 56 | #623 |
| 10 | Qwen 3 30B A3B | 41 | #1122 |
| 11 | Qwen 3 8B | 40 | #1830 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=migue-bench-intra-document-temporal-relations · How It Works · Data refreshed daily, snapshot 2026-11-13.