OpenAI has reset usage quotas for ChatGPT Work and Codex users. The previously paused 5-hour usage limit will be reinstated tomorrow.
Many users noticed that GPT-5.6 Sol burns through quotas faster than expected. The reason isn't a reduction in subscription limits — it's that the model is more willing to work for longer periods and also calls on more tools and sub-agents.
Especially when waiting for tool execution while continuing other tasks, Sol generates more tool calls, responses, and cached tokens. Complex tasks and heavy web searches therefore exhaust quotas more quickly.
OpenAI has optimized the relevant mechanisms. Under normal use, quotas should last about 18% longer.
https://twitter.com/thsottiaux/status/2082317452755751098?s=46