EntLORE - Latent Organizational Reasoning (Agentic Retrieval): 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 latent organizational reasoning questions (L3), whose target relation is stated in no document and must be recovered from routine records, answered through an agent that searches and reads a dense index of the released corpus over at most 30 iterations. Higher is better.
Metric: Answer accuracy (%; L3, 234 questions whose target relation is stated in no document; agentic retrieval: search and fetch tools over a dense index of the released corpus, at most 30 LLM iterations; 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: Around 2029. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 17.5 | #123 |
| 2 | DeepSeek V4 Flash (Reasoning) | 15.1 | #258 (DeepSeek V4 Flash) |
| 3 | Claude Sonnet 4.6 (Thinking) | 14.9 | #161 (Claude Sonnet 4.6) |
| 4 | GPT-5.4 Mini | 14.8 | #389 |
| 5 | DeepSeek V4 Pro (Reasoning) | 14.3 | #188 (DeepSeek V4 Pro) |
| 6 | GLM-5.2 | 12.3 | #138 |
| 7 | Qwen 3.5 397B A17B | 9.5 | #246 |
| 8 | Kimi K2.6 | 9.4 | #178 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=entlore-latent-organizational-reasoning-agentic-retrieval · How It Works · Data refreshed daily, snapshot 2026-11-13.