Which version of GPT-4o does the platform provide?
The available versions are:
However, the platform does not specify which one is being used. I’d like to know which version I’m using because, based on my tests, I’m receiving significantly lower-quality responses compared to when I use the gpt-4o-2024-11-20 version on Dify, for example. Thank you!
We’re using the most recent version. The way we have it set up, it will only use the most recent version.
Hope that helps!
Actually, pickaxe isn’t using the latest version. I ran a controlled test on my API, and pickaxe is using the Claude 3.5 sonnet 2024-06-20 model, even though there’s already a Claude 3.5 sonnet 2024-10-22 model that provides much better answers than the previous one.
I imagine the same is happening with the OpenAI models. As I see the quality of the delivered texts are not that good.
How can we fix this? This is crucial for my business, as the quality of the responses changes significantly.
just switched it over now! Thanks for pointing that out
GPT is still running on the GPT-4o-2024-08-06 version, which is not the latest one. I just tested it
the latest one is 2024-11-20
OpenAI is still on the 2024-08-06. We’re still running QA on the newest weights. Very important to do this.
Claude is on the 10-22-2024 weights.