An Explosive Conspiracy Theory Suggests China’s MiniMax “Sniped” Artificial Analysis’ Evaluation Queries To Make Its Otherwise Shitty M3.1 Flash Model Look Good

Every day brings something new in the AI world, including conspiracy theories that are batshit crazy. One conspiracy theory about MiniMax's new M3.1 Flash model, however, is not crazy at all. In fact, it might be one of the most plausible explanations behind the model's meteoric rise on AI benchmarks, and it all has to do with a circuitous path to the good ol' benchmaxxing, an abhorrent practice where AI labs specifically train their proprietary AI models to ace specific benchmark tests rather than focusing on improving their general abilities. MiniMax allegedly sniped Artificial Analysis' queries during its evaluation of […]