Fine-tuned model in assistant give - Run failed Sorry, something went wrong.

API Bugs
,
21

I’m encountering the same issue with the OpenAI fine-tuned models within assistants. any attempt to use the already fine-tuned model within my assistant results in a “Run failed - Sorry, something went wrong.” error message. However, the same model gpt-3.5-0125 work perfectly when called directly, and there’s no issue when I switch to a standard model like gpt-3.5-0125. the only problem is with the fine tuned model

22

Same issue here, only with finetuned models.

23

Hi,
Has anyone else noticed that the fine-tuned model works perfectly in Playground - Chat? You just select the model, add the prompt to the system, and use it without any issues. It seems very strange to me that it works normally in that section, but doesn’t work in Playground - Assistants. Has anyone else experienced this or has any idea why this might be happening?
Thanks.

24

Exact same issue on our side as well. We are still waiting on this, hoping openai would fix it. No proper response from support yet

25

I understand that more and more people are facing this problem. I ask you to write to tech support, demand that a person answer you, although I am encouraged to call a person and write to the mail, nothing happens, write all together to tech support, maybe soon these inhumans will notice us

26

I have reported the incident to OpenAI’s technical support, and their response has been that they will get back to me in 3 days! :man_facepalming::massage_man:

Hi everyone,
I’m getting the feeling that things are not improving, and it’s causing me a lot of concern for the future. It’s not normal for a paid product to stop working like this without any communication from the developers.
Is anyone working on an alternative? If so, it would be fantastic if you could share it. Thank you very much!

27

I am getting Run failed
Sorry, something went wrong. today as well

28

Same here. This is a VERY bad problem open ai.

29

Did you guys have any update of these problem?

31

Same issue here since yesterday (16.05.2024)
Yo, OpenAI, please don’t let us down on this.
Messaged the support:

Can’t use my fine-tuned gpt-3.5-turbo-0125 model on Assistants. When testing in Playground / Assistants i get only this error:

Run failed Sorry, something went wrong.

Steps to reproduce the error:

  1. OpenAI platform: Fine-tuning > Create
  2. Select the gpt-3.5-turbo-0125 model and upload jsonl file and hit Create
  3. Fine-tuning is complete without any errors
  4. Go to: Assistants > Create
  5. Enter the name of assistant, brief instructions and select the fine tuned model just created (any of three available models with a common ID). Assistant is created
  6. Go to: Playground > Assistans (make sure the created assistant is selected)
  7. Ask any question
  8. The result is:

Run failed Sorry, something went wrong.

3 Likes
32

It started working today! Seems like all my old fine tuned model in assistants are working now.

33

I just tried and is not working yet :smiling_face_with_tear:

34

Same issue as of May 31, 2024. Assistants - Playground is not working with fine-tuned model, returning a “run failed” error without further details

35

Same for me on June 11
As soon as I try to use my fine-tuned model (based on 3.5-turbo-0125) in the playground, the assistant only answers with “Run failed Sorry, something went wrong.”

Still no solution available?

36

Same issue on june 26.

“Run failed Sorry, something went wrong.” when i try to use my fine tuned model based on 3.5-turbo-0125.

When i use a simple chat everything is ok but is not working with Assistant

1 Like
37

Same Issue on Sept 4, also lost the ability to use it on my Assistant API.

38

Getting same issue, model not supported when using finetuned models on Assistants API. Nothing reported on OpenAI outages atm.