Does the Assistants API not support fine tuning?

API
,
1

I tried applying the finetuned gpt3.5 model to my assistant, but my finetuned model does not seem to appear in the list of models on this page https://platform.openai.com/assistants.
image

const assistant = await openai.beta.assistants.create({
  name: "Math Tutor",
  instructions: "You are a personal math tutor. Write and run code to answer math questions.",
  tools: [{ type: "code_interpreter" }],
  model: "my Finetuning model"
});

I don’t think this is valid either.

Does the Assistants api not support finetuning?
If so, I’m also wondering if there are any plans to support finetuning.

1 Like
2

The Assistants API does not support the use of fine-tune models.

The documentation had some contrary information, but this has been confirmed as intended.

It also could be very expensive when your fine-tune emits repeated errors and you can’t see exactly what’s happening. Only 1106 models support parallel tool calls of retrieval functions, and that is already teetering on the edge where fine-tune could damage it.

1 Like