CathedralBench

An outside red team's test: chaining several exploits together to get through a hardened system, scored on the hardest tasks only and run without safeguards, to show what the underlying model can do. Higher is better.

Grok 4.7 by SpaceXAI holds the best CathedralBench result among tracked models, at 29% (released Sep 21 2026). Higher scores are better on this benchmark.

1. Grok 4.7 (SpaceXAI) — 29%; 2. Grok 4.6 (SpaceXAI) — 25%.

2 tracked models have a published CathedralBench score. Scores are collected from published lab reports and benchmark sources, with recorded sources credited alongside the results. We do not re-run the evaluations.