AI Benchmarks Obscure How Far Behind New Models Actually Are | Tech Twitter
AI Benchmarks Obscure How Far Behind New Models Actually Are
Press Space to continue
A lot of swift conclusions are being drawn about Kimi K3 based on fairly saturated benchmarks and ELOs, rather than actually testing it on very hard problems. The AI frontier has already moved so far that a good model that is a still months behind looks like the future to many.