Can't use GPT-4.1-mini fine-tunes in the Assistants API

API Bugs
, ,
1

Hi. I’m building a chatbot using the Assistants API and would like to use a fine-tuned model to improve accuracy.
I noticed that I can use GPT-4o and GPT-3.5-turbo fine-tunes, but, when trying to create an assistant using a GPT-4.1-mini fine-tune with the following code:

await openai.beta.assistants.create({
  name: 'My Assistant',
  instructions: SYSTEM_INSTRUCTIONS,
  model: 'ft:gpt-4.1-mini-2025-04-14:[...]',
  tools: ASSISTANT_TOOLS,
})

I get an error:

The requested model 'ft:gpt-4.1-mini-2025-04-14:[...]' cannot be used with the Assistants API

The fine-tuned GPT-4.1-mini model doesn’t appear as an option in the Assistants Dashboard (https://platform.openai.com/assistants) either:

The image shows a dropdown menu for selecting the model to use in the Assistant. When searching for "4.1", no fine-tunes appear.
The image shows a dropdown menu for selecting the model to use in the Assistant. When searching for "4.1", no fine-tunes appear.771×342 14.5 KB

I find this weird because here, it says that GPT-4.1-mini supports both fine-tuning and assistants API.

Are GPT-4.1-mini fine-tunes simply not supported in the Assistants API? If so, does OpenAI plan to support them in the near future?

1 Like
2

Further: All fine-tuning 4.1 models are simply filtered out of the JSON endpoint that delivers models to the assistants playground.

3

I am also able to reproduce this bug.