Why can't I access the 1M-token model context window when using the Azure API?

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

26.715.12143

What subscription do you have?

Azure PayGo

What platform is your computer?

Windows 11

What issue are you seeing?

The model’s context length should not be entirely controlled by the client. I use Codex because it’s a good product, but the backend I’m connected to is Azure OpenAI, which already supports a 1M-token context window. The client should not hard-code the context length and impose restrictions on my usage. I understand that requests exceeding 272K tokens may incur additional charges, but since I’m using a Pay-As-You-Go subscription, users like us should not be subject to such limitations.

What steps can reproduce the bug?

none

What is the expected behavior?

_No response_

Additional information

_No response_