Leading AI companies are creating new ways to evaluate increasingly sophisticated AI models as current testing methods prove inadequate. According to Cristina Criddle’s report in the Financial Times, companies like OpenAI, Microsoft, Meta, and Anthropic are developing internal benchmarks because their latest AI systems achieve over 90% accuracy on existing public tests. Meta’s generative AI lead Ahmad Al-Dahle notes that measuring these advanced systems has become increasingly difficult. The industry faces challenges in establishing standardized evaluation methods for AI agents capable of autonomous task execution, with experts particularly focused on testing reasoning abilities. New public benchmarks like SWE-bench Verified and FrontierMath are emerging, but researchers warn that without agreed-upon standards, comparing different models’ capabilities remains problematic.