Hello NVIDIA Team,
I am using a Personal organization on NVIDIA Build.
My active API key authenticates successfully against:
GET https://integrate.api.nvidia.com/v1/models
HTTP 200
However, inference requests to:
POST https://integrate.api.nvidia.com/v1/chat/completions
consistently return:
HTTP 401 Unauthorized
Authentication failed
I reproduced this directly against NVIDIA’s endpoint, without LiteLLM or any third-party proxy.
I tested multiple chat-capable models, including:
- ai21labs/jamba-1.5-large-instruct
- nvidia/nemotron-3-super-120b-a12b
The behavior is the same for both.
Could you please check whether the “Public API Endpoints” entitlement is enabled for my Personal organization and enable it if necessary?
Account type: Personal / Free Tier
Use case: personal development and testing with NVIDIA NIM hosted APIs.
I can provide the Key ID privately if needed.
I will not post the raw API key.
Thank you.