Hello NVIDIA Team,
I am developing an AI companion application called LAHLAH and testing the NVIDIA Nemotron API.
My API key successfully authenticates with the models endpoint:
GET /v1/models — HTTP 200
However, chat completion requests consistently fail:
POST /v1/chat/completions — HTTP 403 Forbidden
Error:
Authorization failed
I tested these models:
-
nvidia/nemotron-3.5-lightning-30b-a3b
-
nvidia/nemotron-3.5-bolt-e4b
Both models appear in my accessible model catalogue, but inference requests return the same 403 error.
Could you please investigate whether my account or organization has the necessary hosted inference permissions enabled?
Thank you for your assistance.