Comment by dev_l1x_be

Benchmarking is grossly misleading. Claude’s subscription with Code would not score this high on the benchmarks because how they lobotomized agentic coding.