BlogLLM Benchmark Scores as Published, Side by Side
CathedralBench: Results as Published
Every result the catalog holds for CathedralBench: 2 models from 1 developer, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area CybersecurityData version 2026103001, as of 2026-10-30
CathedralBench is described by one third-party tracker as a cyber challenge in which an agent must chain several exploits to get through a hardened system. Results are reported only on the hardest tasks, and the runs are without safeguards, to show what the underlying model can do when it is not restricted. This description is secondhand and was not checked against a primary source. No scores are given.
Not checked against a maintainer's page.
CathedralBench (hard subset)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Grok 4.7xAI, released 2026-09-21 | 29% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: xhigh |
| Grok 4.6xAI, released 2026-08-12 | 25% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.