Model fallback and reliability: what happens per request

To keep chat working at peak times, we may maintain multiple fallback routes. If one path is briefly unavailable, we try the next—ideally without you noticing.

You might still see rate limits or regional differences depending on upstream providers. If failures persist, retry later or switch to Basic’s free pool to validate network and session state.

Usage and billing reflect the model that actually served the request, within your plan rules (e.g., wallet deductions).