Possible to use with Openrouter.ai?

[Jirito0]

I'd like to use this with the models supported by Openrouter- the GPT4 32k in particular to help tackle the token limit for the time being.

Openrouter supports apps that use OAI API keys, and I attempted inserting my Openrouter key, and stated in the OAI Config List to use GPT4-32k, but I just get this error: "OpenAI.error.AuthenticationError: Incorrect API key provided: sk-or-v1-83f***********"

I also reviewed the sections regarding using Local LLMs, but it doesn't provide any information regarding using API's like Openrouter.

[sonichi]

I don't have experience with it. I asked on Discord.

[vanetreg]

I'd like to use this with the models supported by Openrouter

Hey @Jirito0
did it work finally? Could you share details?
I plan to try it now, either directly or via LiteLLM proxy...

[sonichi]

@ekzhu @olgavrou @BeibinLi fyi

[reevolutiva-it]

It works for me.
{
"model": "mistralai/mixtral-8x7b-instruct",
"base_url": "https://openrouter.ai/api/v1",
"api_key": "sk-or-v1-***********************************"
},

[yingdaomayi-qwe]

Hey, @reevolutiva-it
where should I add this code?
Thank you!