48-hour test: GPT-6 Sol on $200 plan and Opus 5.5 on $20 each burn only ~8% weekly limits

i_dg23 · x · 2026-09-25

A developer shared usage numbers after nearly 48 hours of heavy use of GPT-6 Sol and Claude Opus 5.5:

The author praises OpenAI's work on Sol and notes Anthropic improved value via limit resets and higher usage caps. His workflow: let the two models collaborate — one builds, the other challenges and improves the output.

Original post →

More from Models

Models channel →