A16z-Backed Vals Aims to Modernize AI Benchmarking with $40M Series A

Benchmarking has long served as the tech industry's primary method for validating AI model capabilities and driving public relations. However, legacy benchmarks have increasingly struggled to keep pace, allowing companies to engineer workarounds for tests that were never designed for modern frontier models. Startup Vals, founded in 2024, is working to overhaul this evaluation framework. Following rapid adoption and an initial seed round backed by 8VC and Bloomberg Beta, the company secured $40 million in a Series A funding round led by venture firm Andreessen Horowitz. Co-founder Rayan Krishnan, a 25-year-old Stanford alumnus with previous experience at Palantir, Microsoft, and the Stanford Artificial Intelligence Laboratory, noted that academic benchmarks have lagged behind commercial frontier advances. Operating out of San Francisco, Vals focuses on building real-world evaluations to verify whether cutting-edge models truly perform as advertised.