GPT-3.5 Incorrectly being marked as disabled in conversations not utilizing tools.

Question

Frequently, ever since the release of GPT-4o, I have noticed that GPT-3.5 is disabled incorrectly even in conversations that do not utilize tools.

For example, I may create a new conversation that is solely GPT-3.5, and it may randomly state that I have utilized tools in the conversation, disabling the conversation, and leading to a disconnect in engagement as a user.

As mentioned, I began seeing the issue when GPT-4o was released. I attempted to remedy the issue through tactics such as clearing my memory and even fully deleting all my conversations. These attempts have been ineffective.

I have been unable to find a correlation between the disablement of the chat as a result of actions before this output. It should also be noted that the output can be produced even when new queries are not generated, but merely when I use the "Retry message" feature.

Thank you for the help!