What happened?

Setting allowed_openai_params: ["tools"] under

- model_name: model
  litellm_params:
    allowed_openai_params: ["tools"]

is not respected in requests. The only way I could get it working was attaching "allowed_openai_params": ["tools"] in the body of my requests. It should be enabled by default with this config option

Relevant log output

litellm.UnsupportedParamsError: together_ai does not support parameters: ['tools'], for model=XYZ To drop these, set `litellm.drop_params=True` or for proxy:\n\n`litellm_settings:\n drop_params: true`\n. \n If you want to use these params dynamically send allowed_openai_params=['tools'] in your request..

The model I'm testing with is a TogetherAI model with a custom name, is it possible this is related to the bug? Do custom models need to be explicitly given support somewhere else in the config?

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

v1.77.3-stable

Twitter / LinkedIn details

No response