The max prompt tokens always show 128K

[ricoisme]

Does this issue occur when all extensions are disabled?: No

• VS Code Version: 1.109.0
• OS Version: Windows

It seems max prompt tokens are always fixed at 128k, even When I switch to the GPT-5.2-Code model.
Is this expected behavior?

Steps to Reproduce:

1. Use copilot cloud agent with GPT-5.2-Code model
2. Send a prompt and receive an error message within a few seconds.

Error Failed to get response from the AI model; retried 5 times (total retry wait time: 5.849579995864053 seconds) (Request-ID F019:2195FF:2FEC7CB:370DA9B:6985E7DE) Last error: CAPIError: 400 prompt token count of 184399 exceeds the limit of 128000