Chat response terminates early | Mistral AI | Page 1

jaunty garnet Jun 17, 2025, 10:22 PM

#

I'm suddenly getting an 400 API error: "Fine-tuning is not enabled for this model."

But this is not the case. I've been using this model for nearly a year without issue. This has happened out of the blue without any changes to my application.

pliant needle Jun 17, 2025, 10:26 PM

#

Hi @jaunty garnet , which model exactly?

jaunty garnet Jun 17, 2025, 10:35 PM

#

ft:mistral-small-latest:b9bbf274:20240610:54d3f457

pliant needle Jun 17, 2025, 11:07 PM

#

@jaunty garnet im not sure I'm understanding, you are getting the error when using the model or when finetuning, since the error you seem to get is for finetuning

jaunty garnet Jun 17, 2025, 11:09 PM

#

When using the model.

#

I create a new fine-tuned model and am using that successfully.

#

I got an email a few weeks ago saying mistral-small-2402 would be deprecated but I didn't think this applied to me. The Fine-tuned modle I was using was based on mistral-small-latest created 7/18/2024

pliant needle Jun 18, 2025, 11:39 AM

#

Feel free to ask for help via the support button directly https://help.mistral.ai/en/
There is a button in the bottom right, you can also access it via the plateforme and le chat in the bottom right corner

Mistral AI - Help Center

split egret Jun 18, 2025, 12:29 PM

#

Hello @jaunty garnet this is an oversight of our communication, we are very sorry for that. The base model for your finetune is indeed mistral-small-2402 it was removed from our inference API a few days ago which explains it no longer works.

I added 25USD in your account so you can retrain a model on the new small.

We are making two changes in the next few days:

Depreciation of fine-tuned models will be more explicit and shared via email ahead of time
The root version for fine-tuned models will also be more visible rather than the -latest alias

#Chat response terminates early