Custom model provider still consuming AI quota

[0xMaluco]

When using an alternative model provider with GitHub Copilot Desktop, the AI quota/limit is still being consumed from my account instead of using the custom provider's quota.

Expected behavior:

• Using a custom model provider should consume that provider's tokens/quota, not my GitHub AI quota

Actual behavior:

• Even with a different model provider configured, my GitHub AI account quota is still being decremented

Steps to reproduce:

1. Configure a custom model provider in Copilot Desktop settings
2. Use the custom provider for code completions
3. Check account quota - it's still being used

This appears to be a bug where model provider selection doesn't properly isolate quota consumption.