🦞Clawmark™ Verified

DAWES™ Benchmark

Domain-Aware Weighted Evaluation System — measuring real-world AI capability in insurance & engineering scenarios.

Leaderboard

Loading benchmark data…

Benchmark Archive

v1.0 — 2026-04-13 · 13 models
Initial benchmark — 60 domain-specific I&E questions, 13 models evaluated.
↓ Download JSONSubmit to Scrutiny →

Papers & Methodology

📄DAWES Benchmark v1 Paper✉️Open Letter to Trades🛡️Carapace Protocol White Paper🔬Methodology Document

Are you a lab or researcher?

We welcome scrutiny. Reproduce our results, prove us wrong, or extend the benchmark. All data is open.

View on GitHub →