I’m seeing a reproducible Pro entitlement / usage-limit issue in the California State University (CSU) ChatGPT Edu environment.
Pro has remained completely disabled and cannot be selected, so no additional Pro usage has been possible since the issue began.
However, the displayed reset date moves forward by one day every day:
- September 3, 2026: “Try again after October 3”
- September 4, 2026: “Try again after October 4”
- September 5, 2026: “Try again after October 5”
The reset date therefore appears to move forward continuously instead of remaining at a fixed monthly reset date.
Expected behavior:
If Pro usage has reached a monthly limit, the displayed reset date should remain fixed until that usage period resets.
Actual behavior:
Pro remains disabled, but the reset date advances by one day every day even though no Pro usage is possible.
Additional observations:
- The affected account is on a managed ChatGPT Edu workspace.
- The issue occurs in both the ChatGPT desktop app and web interface.
- GPT-5.6 Sol High is still working normally on the web.
- A clean browser-network test with High selected showed:
model: gpt-5-6-thinking
thinking_effort: extended
resolved/server model: gpt-5-6-thinking - This suggests the issue may be specific to Pro entitlement, quota accounting, or reset-date handling rather than a complete loss of reasoning-model access.
- The behavior started during the recent managed-workspace/model rollout period.
Reproduction steps:
- Open ChatGPT in a managed Edu workspace.
- Open the model picker.
- Observe that Pro is disabled.
- Hover over or inspect the Pro limit message and note the reset date.
- Do not use Pro.
- Check again the following day.
- The reset date advances by one day.
Has anyone else on an Enterprise or Edu managed workspace seen the same rolling reset-date behavior?
I have screenshots from multiple consecutive days, but workspace/account identifiers have been redacted for privacy.
Related model availability issue
This CSU Edu environment also still does not expose the Extra High reasoning option, despite OpenAI’s Enterprise/Edu documentation listing Extra High as part of the managed-workspace model picker, subject to rollout and workspace access settings.
This managed Edu workspace also still does not have Extra High, despite the June 26 Enterprise/Edu release notes listing Extra High as one of the available model-picker options for Enterprise and Edu workspaces, subject to workspace access settings.
Could OpenAI clarify whether Extra High is still being rolled out to Edu, requires an additional admin setting, or has other eligibility requirements?
If this is a delayed rollout, I would also like to request broader Extra High availability for eligible Edu workspaces.
Recent managed-workspace model changes
With GPT-6 Astra and other model-access changes now rolling out, clearer documentation about model availability, entitlement changes, and rollout timelines for Edu workspaces would be very helpful.
