Benchmarks

Rigorous, reproducible LLM benchmarks with open datasets and documented methodology. Every result links to raw CSV data so you can verify, extend, or challenge our findings.