informationisbeautiful.net
A visualisation of major large-language models ranked by performance using the MMLU benchmark, which consists of 16,000 multiple-choice questions across 57 academic subjects. It also presents selected lawsuits filed against AI companies and notes that MMLU has some critiques.