4o mini thinks it's a GPT 3 model?

I’m trying 4o-mini and I’m asking it what model it is–it seems to think it’s a GPT 3 model through the API. Is anyone else experiencing this?

from openai import OpenAI
client = OpenAI()

completion = client.chat.completions.create(
  model="gpt-4o-mini",
  messages=[
    {"role": "user", "content": "What model are you?"}
  ]
)

print(completion.choices[0].message.content)

Results in

I am based on OpenAI's GPT-3 model, specifically designed to assist with a wide range of questions and tasks. If you have any specific queries or need assistance, feel free to ask!

As always, don’t ask models about themselves, they can’t know what they are without being told.

1 Like