Vals, backed by Andreessen Horowitz, is looking to become the gold standard for AI benchmarking

Vals is a startup that tests artificial intelligence models to measure their real-world capabilities across industries like law, finance, and coding. The company matters because AI benchmarking has become an industry standard for validating model performance, but existing benchmarks are outdated and companies have learned to game them by training their models against publicly available tests. Vals differentiates itself by keeping its test materials private and evaluating whether models can complete complex tasks at human-level quality rather than just assessing general knowledge. The startup has grown rapidly after raising $40 million in Series A funding and now works with companies that pay to have their models evaluated, a revenue model the founder compares to students paying to take standardized tests.