I would like to report three distinct issues.
- Codex usage reset
I am confident that this issue is reproducible.
Whenever an official Codex usage reset is applied while an account is actively using Codex and consuming quota, that account does not appear to receive the reset. Based on my observations, this affects multiple users, not merely my own account.
I have encountered this several times, which leads me to believe there may be a flaw in the way resets are applied to accounts that are active at the moment the reset is issued.
- ChatGPT model routing
This issue may be specific to my account, so I am posting it here to determine whether anyone else has experienced similar behavior.
I am a 5x subscriber, yet my requests do not appear to be routed to the model I selected.
When I select Pro, my requests appear to be routed to GPT-5.5 mini.
For GPT-5.6 Sol, I cannot determine which backend model is actually being used.
As a simple test, I sent only “Hi,” and immediately reached the usage limit.
I am unsure whether this is expected behavior, an account-level issue, or a routing bug. If anyone else has observed the same behavior, I would appreciate hearing about it.
- Codex update: unsupported parameter error
After updating Codex, I began receiving the following error whenever I attempted to send a message:
“Unsupported parameter: ‘reasoning.summary’ is not supported with the ‘gpt-5.3-codex-spark’ model.”
I am not calling the API directly, and this issue did not occur before the update. It appears that the updated Codex client is sending a parameter that the ‘gpt-5.3-codex-spark’ model does not support.
This may indicate a client-side regression or a model-routing/configuration issue. Has anyone else encountered the same error after updating Codex?
