About
Why we’re building Byggeriets AI Benchmark
Byggeriets AI Benchmark is an independent research initiative measuring how reliably large language models can interpret and apply the Danish Building Regulations.
Mission
Our mission
Danish building regulations are dense, frequently updated, and carry real safety and legal consequences when misread.
We publish our benchmark questions, our scoring methodology, and our full results in the open.
The project is led by a small, interdisciplinary team based at UCN, working in collaboration with TUM.
Team
Who’s behind the project
A small, interdisciplinary team investigating how AI performs on real regulatory judgment calls.
Based at UCN, in collaboration with TUM.
Partners
Contact
Get in touch
Questions about the benchmark, the dataset, or a potential collaboration — we’d like to hear from you.