Any comparison that brings "frontier-like" claims should have benchmarks against Opus 4.8 (or Fable 5 once we get it back) and GPT 5.5. Spoiler alert: this art

Any comparison that brings "frontier-like" claims should have benchmarks against Opus 4.8 (or Fable 5 once we get it back) and GPT 5.5. Spoiler alert: this article just says that GLM 5.2 is better in quality than MiniMax M3, but worse value for money.