mirror of
https://github.com/vercel-labs/fx.git
synced 2026-10-28 18:25:19 +08:00
Pranit
09ec38c243
## Summary - Send chat requests to the v4 AI Gateway language-model endpoint so reasoning effort settings are applied instead of dropped by the v3 route. - Send an effort of "max" as "xhigh", the highest tier the v4 endpoint accepts, so models whose catalogs advertise "max" no longer fail with HTTP 400.