Plus users say GPT-6.1 Sol started eating usage limits much faster on Oct 16 — OpenAI hasn't confirmed

#What happened

On Oct 17, a Plus subscriber posted on OpenAI's community forum that GPT-6.1 Sol began consuming their Plus allowance "dramatically faster" starting Oct 16 — same model, same settings — and asked OpenAI whether Sol's usage weighting or the included allowance had changed. Over the next day the thread grew:

• joswak reported a "very noticeable shift in consumption" on Oct 18.
• SemperFi said the same workloads on the same model suddenly exhausted their allowance within five hours, and spent three hours trying to get the issue escalated past automated support to a human investigator.
• harryrx8 reported roughly double consumption for similar tasks since yesterday, and asked OpenAI to disclose whether it was a deliberate policy change.
• uruzu, on a Pro 200 plan, found Sol High used 51% of a reset overnight but only 6% the next morning for the same Excel-processing task.
• Ni-Ni estimated a 4-5x faster drain and noted the issue also appears on GPT-6 Luna, suggesting it is not isolated to Sol.

#Why it matters

GPT-6.1 Sol's selling point over Astra is cost efficiency (one-fifth the API price), so a silent change in subscription usage accounting would erase that advantage for paid users. The reports also overlap confusingly with Tibo's 28-day "Day 13 limit reset" campaign, and one user noted the allowance can decrease even with no visible task running. There is no confirmation from OpenAI yet: no official metering change has been announced and no fix has been promised.

#Sources

• OpenAI community forum thread: GPT-6.1 Sol suddenly consuming much more Plus usage since October 16

Verified 2026-10-18.