Not really, it's just that our benchmarks are not good at showing how they've improved. Those that regularly try out LLMs can attest to major improvements in re

Not really, it's just that our benchmarks are not good at showing how they've improved. Those that regularly try out LLMs can attest to major improvements in reliability over the past year.