GPT-5.3-Codex-Spark quota is 100% but Spark is missing from model selector and Codex remains blocked

What version of the Codex App are you using (From “About Codex” dialog)?

26.909.12148

What subscription do you have?

ChatGPT Pro

What platform is your computer?

Windows x64

What issue are you seeing?

I am unable to use GPT-5.3-Codex-Spark in the Codex Windows app even though my account shows that the dedicated Spark quota is fully available.

Environment

• Codex App version/runtime: 26.909.12148
• Subscription: ChatGPT Pro
• Platform: Windows x64

Current usage status

The Usage page currently shows:

• General Codex/Work weekly limit: 0% remaining
• General weekly reset: September 19, 2026 at 20:37
• GPT-5.3-Codex-Spark 5-hour limit: 100% remaining
• GPT-5.3-Codex-Spark weekly limit: 100% remaining

The Spark 5-hour quota had already been exhausted earlier, but more than 5 hours have passed and the Usage page now clearly shows that it has reset to 100%.

Actual behavior

Codex is still blocked and displays:

"You’re out of Codex and Work usage.
Add credits to keep going now, or wait for usage to reset on Sep 19 at 20:37."

In addition, GPT-5.3-Codex-Spark does not appear in the model selector at all.

The models currently available in my model selector are:

• GPT-6 Astra
• GPT-5.6 Sol
• GPT-5.6 Terra
• GPT-5.6 Luna
• GPT-5.5

GPT-5.3-Codex-Spark is missing.

Expected behavior

Because my account shows:

• GPT-5.3-Codex-Spark 5-hour quota: 100% remaining
• GPT-5.3-Codex-Spark weekly quota: 100% remaining

I expect GPT-5.3-Codex-Spark to appear in the model selector and to remain usable while its dedicated quota is available, even though my general Codex/Work weekly quota is exhausted.

Current inconsistent state

• General Codex/Work weekly quota: 0%
• Spark 5-hour quota: 100%
• Spark weekly quota: 100%
• Spark model: missing from model selector
• Codex input: globally blocked until Sep 19

This appears to be an entitlement, model availability, or quota-gating synchronization issue.

Please investigate whether GPT-5.3-Codex-Spark entitlement/model availability is out of sync with the quota state on my account.

What steps can reproduce the bug?

Feedback ID: no-active-thread-01a0a3f9-1a90-7f40-8527-1bd15be4926c

What is the expected behavior?

No response

Additional information

No response

Start by reproducing the issue in the Windows x64 Codex app using the reported version 26.909.12148, then compare the Usage page, model selector, and Codex blocking state using feedback ID no-active-thread-01a0a3f9-1a90-7f40-8527-1bd15be4926c. Done means identifying why the Spark entitlement and quota state diverge, with Spark selectable while its dedicated quota remains available.