Doc2DB-Bench - Global Aggregation: leaderboard
Metric: Cell-level F1 (%; relation-table cells annotated with the Global Aggregation 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 | 82.2 |
| 2 | Gemini 2.5 Pro | 73.03 |
| 3 | GPT-4o | 71.59 |
| 4 | Claude Opus 4.6 | 64.25 |
| 5 | Qwen 3 Max | 62.15 |
| 6 | DeepSeek V4 Flash | 59.83 |
| 7 | Qwen 2.5 14B Instruct | 49.23 |
| 8 | Llama 3.1 70B Instruct | 48.3 |
Interactive version: theaggregate.ai/benchmark?slug=doc2db-bench-global-aggregation · How It Works · Data refreshed daily, snapshot 2026-09-29.