Some OpenAI models are not visible

[neelabalan]

Some models like o3-mini and o4-mini are not visible. I tried this with my API key which I provided in the interface and I have access to these models.
curl https://api.openai.com/v1/models -H "Authorization: Bearer <>

I tried them from the OpenAI playground too. I checked out the commit from main branch (0103b4b08a6ee13) which is version 0.7.8. Followed instructions form here. I'm not sure if I'm missing anything. Help appreicated.

[DoS007]

Similar problem here.

The librechat.yaml in hugging face is outdated it seems. Interestingly sonnet 4 is shown although not configured either in the yaml.

I think it would be best to remove the whole model names in the .yaml and just to collect the current models via API and store the results in database so that the db is gradually updated if new models show up (you could also make a "refresh models" button or similar).
Or at least to that additionally to the yaml (at least if a appropriate key is present).

My cfg was 1.2.4 and on github the example is even older with 1.2.1 https://github.com/danny-avila/LibreChat/blob/main/librechat.example.yaml .

The newest cfg is available here https://www.librechat.ai/docs/configuration/librechat_yaml/example with changelog here https://www.librechat.ai/changelog .

---

I copied the example from the docs (1.2.8, newest version) and also done factory-reset for updating in hugging face. Also provided openai key (can use chatgpt latest). Still e.g. no openai 4.1 on models.

[neelabalan]

Updated my yaml file. Thanks.

[danny-avila]

All models for the HF space are manually added by the admin, I'll make a note so they can update it.

[DoS007]

Did factory reset for updating: Still no o3 and 4.1 and so on.

[DoS007]

Maybe found a solution, see below.

[DoS007]

Maybe found a solution: Edit the variable OPENAI_MODELS on hugging face and append it with ",o3,o1".