Author Calls GPT 5.6 Sol the Weakest Yet

teodorio · x · 2026-07-08

After testing GPT 5.6 Sol for months, the author claims it is one of the weakest models they’ve used, with capabilities falling below even 5.5 pro. They speculate that AGI progress has been delayed and dismiss related equity packages as "worthless."

Original post →

More from Models

Models channel →