BizBench - SEC-Num: leaderboard
Metric: Accuracy (%; 2,000 SEC 10-K and 10-Q filing snippets, extract the quantity matching an XBRL label, 3-shot). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 | 79.3 |
| 2 | GPT-3.5 | 76 |
| 3 | Llama 2 70B | 74.7 |
| 4 | Mixtral 8x7B | 73.1 |
| 5 | Mistral 7B | 65.1 |
| 6 | Llama 2 13B | 61.4 |
| 7 | starcoder | 57.8 |
| 8 | mpt-30B | 56.4 |
| 9 | Llama 2 7B | 56.2 |
| 10 | falcon-40B | 52.3 |
| 11 | mpt-7B | 31.7 |
| 12 | falcon-7B | 27.3 |
Interactive version: theaggregate.ai/benchmark?slug=bizbench-sec-num · How It Works · Data refreshed daily, snapshot 2026-09-26.