Herculean - Auditing (OpenClaw): leaderboard
Metric: Accuracy (%; share of the 65 SEC XBRL audit instances where the agent both extracts the reported value of the target concept and computes its correct value from the filing's calculation network, graded by the FinAuditing hierarchical LLM judge; OpenClaw agent with this backbone, tools through the Herculean auditing MCP server). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 (OpenClaw) | 66.15 |
| 2 | GPT-5.4 (OpenClaw) | 66.15 |
| 3 | Qwen3.5-397B-A17B (OpenClaw) | 43.08 |
| 4 | Qwen3.5-27B (OpenClaw) | 36.92 |
Interactive version: theaggregate.ai/benchmark?slug=herculean-auditing-openclaw · How It Works · Data refreshed daily, snapshot 2026-09-26.