EntLORE - Cross-Source Composition (Gold Documents): leaderboard
EntLORE (2026) builds an anonymized enterprise world of 2,341 routine documents from an audited truth graph. This board is answer accuracy in percent on its cross-source composition questions (L2), which combine facts stated in several sources, answered through an agent that is handed the gold document paths and reads them through the same tool loop. Higher is better.
Metric: Answer accuracy (%; L2, 204 questions composing facts stated in several sources; gold-document agentic reference: the gold document paths read through the same agent tool loop; 907 questions over 2,341 anonymized enterprise documents; programmatic scoring for entity, set, count and ordered answers, atomic-claim entailment judged by Claude Opus 4.8 for free-form answers). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 397B A17B | 95.9 | #246 |
| 2 | GLM-5.2 | 95.3 | #138 |
| 3 | GPT-5.4 | 95.2 | #123 |
| 4 | DeepSeek V4 Pro (Reasoning) | 95.2 | #188 (DeepSeek V4 Pro) |
| 5 | Kimi K2.6 | 94.9 | #178 |
| 6 | DeepSeek V4 Flash (Reasoning) | 94.7 | #258 (DeepSeek V4 Flash) |
| 7 | Claude Sonnet 4.6 (Thinking) | 94.5 | #161 (Claude Sonnet 4.6) |
| 8 | GPT-5.4 Mini | 84.9 | #389 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=entlore-cross-source-composition-gold-documents · How It Works · Data refreshed daily, snapshot 2026-11-13.