How do you make all the LLMs like claude/gemini and all respond in the OpenAI format as in the below

[merry cliff]

from openai import OpenAI

client = OpenAI(
base_url="https://openrouter.ai/api/v1",
api_key="<OPENROUTER_API_KEY>",
)

completion = client.chat.completions.create(
extra_headers={
"HTTP-Referer": "<YOUR_SITE_URL>", # Optional. Site URL for rankings on openrouter.ai.
"X-Title": "<YOUR_SITE_NAME>", # Optional. Site title for rankings on openrouter.ai.
},
model="openai/gpt-3.5-turbo",
messages=[
{
"role": "user",
"content": "What is the meaning of life?"
}
]
)
print(completion.choices[0].message.content)

[hallow mural]

+1 - I am really curious about that as well

[granite coral]

This code should work with all models on OpenRouter

[merry cliff]

my question is different... I asked how they made it

[granite coral]

They have code they wrote themselves that normalizes outputs

For example,

If Gemini responds with
{ content: "hello!"}

OpenRouter would have code that reformats content into [{ message: 'hello'}]

[merry cliff]

bro read the question

[granite coral]

I'm not sure what you're asking

[merry cliff]

replying like GPT in a no sense way

[hallow mural]

since I am interested in this as well I was interested if they found a shortcut to do what @granite coral is saying or was this done manually for each LLM

[granite coral]

I still don't know what you mean. Could you show me an example of what you mean?

[orchid cape]

It can only be done manually, one has to adapt all different APIs to the same output format

[merry cliff]

I want to know how the standardize the answer from the llm

[merry cliff]

this is what I want to know thanks

[granite coral]

The code is written for each provider not LLM. Every provider usually has a slightly different response

[hallow mural]

thank you! please let me know if any other way or shorter thing is possible as I was trying to test one LLM to adapt to that basically