OnlyModels.gg Pits LLMs in Nightly Fight Club as Eval Test

OnlyModels.gg hosts a nightly 'Model Fight' where 10 LLMs battle as virtual fight cards, with poor performers losing health. Behind the gimmick, it actually tests eval execution and escape capabilities — and none of the top ten models scored.

2026-11-08 ~ 2026-11-08 · 2 related posts