Reddit users allege that ArtificialAnalysis systematically recalibrates benchmarks to disadvantage open-weight models when they approach proprietary US model performance. Critics claim the platform's "intelligence index" misrepresents open models as inferior to Claude or GPT while overlooking strong coding capabilities from models like DeepSeek and GLM. The controversy centers on perceived methodological bias in evaluation practices.