// TAG: AI BENCHMARKS

While proprietary LLMs still hold a slight edge in peak performance for specific tasks, open-weight models have dramatically closed the gap by …