Doc2DB-Bench - Mutual Exclusion: leaderboard
Metric: Cell-level F1 (%; relation-table cells annotated with the Mutual Exclusion capability of the inter-table taxonomy (Section 5.3); 203 synthesized long-document instances over 42 database schemas from BIRD and Spider in seven domain groups; identical prompts, greedy decoding at temperature 0; cells aligned by global maximum-weight tuple matching, a cell matching on exact numeric equality or at least 90 percent string similarity). Source: arxiv.org. Saturation forecast: Around December 2026. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 74.46 |
| 2 | Gemini 2.5 Pro | 70.39 |
| 3 | Claude Opus 4.6 | 61.96 |
| 4 | GPT-4o | 60.18 |
| 5 | DeepSeek V4 Flash | 58.45 |
| 6 | Llama 3.1 70B Instruct | 56.94 |
| 7 | Qwen 3 Max | 54.61 |
| 8 | Qwen 2.5 14B Instruct | 36.41 |
Interactive version: theaggregate.ai/benchmark?slug=doc2db-bench-mutual-exclusion · How It Works · Data refreshed daily, snapshot 2026-09-29.