AI NewsOpenAI Researcher Says AI Benchmarks Fail to Measure Frontier Model Capabilities

OpenAI Researcher Says AI Benchmarks Fail to Measure Frontier Model Capabilities

11:33 AM IST · June 10, 2026

Noam Brown argues benchmark scores increasingly depend on inference compute, obscuring the true capabilities of frontier models.