Hi NVIDIA Team,
I’m using the NVIDIA API Catalog (build.nvidia.com) with an OpenAI-compatible
client. My API key is valid — GET /v1/models returns the model list successfully.
However, when I try to call the chat completions endpoints for the free/public
models (POST /v1/chat/completions), the request fails and I cannot access any
of the free models.
I believe my account is missing the “Public API Endpoints” permission. Could you
please enable it for my account (Personal) so that I can use all the free/public
models available in the NVIDIA API Catalog?
Account Info
- NVIDIA account email: 1sonder9527@outlook.com
- Organization / Personal Org: Personal
- API Key ID (visible prefix): 07227f21-77a1-4da5-…
(actual nvapi- key prefix if available: nvapi-XXXX…)
Verification I’ve Done
- ✅ GET /v1/models → 200, returns model list
- ❌ POST /v1/chat/completions → fails / cannot access free models
(I no longer have the exact error code on hand, but the call does not
succeed for any of the free models)
What I’m Trying To Do
Access and use all free/public models in the NVIDIA API Catalog via the
OpenAI-compatible API.
Request Body Example
{
“model”: “meta/llama-3.1-8b-instruct”,
“messages”: [{“role”: “user”, “content”: “hi”}],
“max_tokens”: 50
}
Thank you!