Labs self-report benchmark scores 6.4% higher, study finds

Open-source leaderboard whatmodel.sh analyzed 1,629 published benchmark results, matching 611 comparisons of 180 models across 103 benchmarks. It found labs reported scores 6.4% higher on average than independent testers, with data and code released openly.

• China's Open-Weight Models Trail US Closed Models by 7 Months2026-10-14