Defending GPT-6.2 Astra: beats Fable 5.2 at chess, produces novel math proofs

VraserX · x · 2026-11-19

VraserX pushes back against influencers calling GPT-6.2 Astra the worst model ever, citing its chess win over Fable 5.2 and proofs for previously unformalized math problems, arguing critics should test the model themselves instead of parroting takes.

Original post →

More from Models

Models channel →