gpt-6.1-sol grinds 35+ minutes on trivial validation prompt at xhigh setting

arthurcolle · x · 2026-09-30

Developer arthurcolle reports gpt-6.1-sol at xhigh reasoning spent 35+ minutes and counting on a trivial prompt to validate a single line of code. He jokes about betting whether it would run until exhausting his usage quota — a real cost concern for paid users given aggressive reasoning budgets.

Original post →

More from Models

Models channel →