AI NEWS

GPT-5.5, Model Cards, And Trustworthy Evaluation

Comparing early GPT-5.5 performance to Claude Opus and arguing for cross-lab safety testing beyond vibes.

Joseph · Apr 28, 2026

0 0