Hi NVIDIA team, I'm getting a persistent HTTP 403 "Authorization failed" when calling the chat completions endpoint, while the models endpoint works

AI & Data Science NVIDIA NIM Access/Accounts
1

Hi NVIDIA team,

I’m getting a persistent HTTP 403 “Authorization failed” when calling the
chat completions endpoint, while the models endpoint works fine with the
very same API key.

What works:
GET https://integrate.api.nvidia.com/v1/models
→ HTTP 200, returns the full model list. The key authenticates correctly.

What fails:
POST https://integrate.api.nvidia.com/v1/chat/completions
→ HTTP 403
→ {“status”:403,“title”:“Forbidden”,“detail”:“Authorization failed”}

This happens with every model I have tried, including openai/gpt-oss-20b.

Account details:
Account email : tum28030@gmail.com
Organization : Personal
API key name : Arko-lucia
API key prefix: nvapi-xZpwvuPgtMj5
Key status : Active

I have already tried regenerating the key; the behaviour is identical, so
this does not look like a key problem. Based on other threads in this
category with the exact same symptom (models list works, completions
return 403), it appears my personal organization is missing the
“Public API Endpoints” permission.

Could you please enable “Public API Endpoints” for my personal organization?

I am happy to provide any further account details, but I will not post the
raw API key.

Thank you.

2

Hello NVIDIA team,

I am getting HTTP 403 Authorization failed when calling:
POST https://integrate.api.nvidia.com/v1/chat/completions

However, my API key successfully works with:
GET https://integrate.api.nvidia.com/v1/models

I tested multiple models and they all return the same 403 error.

My API key is active, but I cannot find the Public API Endpoints permission in my personal organization/account.

Could you please check my account and enable/restore access to the Public API Endpoints for hosted NIM inference?

This is for personal development and testing.

Thank you.