OpenAI has silently rolled out the GPT-4o model as the default for all GPTs. The new model, GPT-4o, is noted for being fast but also verbose, and there are concerns that it may not perform as well as the previous GPT-4 model. Users have reported that GPT-4o is not better than GPT-4, particularly for challenging use cases. Custom GPTs are already using GPT-4o. This change may require users to update their GPT instructions to adapt to the new model's verbose capabilities.
i'm getting the impression that for a lot of challenging use cases gpt4o is actually not better than gpt4. what's your experience?
Yes, GPTs now use GPT-4o. OpenAI has silently rolled out the GPT-4o model as the default for all GPTs.GPT-4o is fast but also verbose and definitely not better than GPT-4, so you might have to update your GPT instructions now as it may not be as good as before . The worst part… https://t.co/7PkICv7fGs https://t.co/FMWVP2ykLi
Custom GPTs are already using GPT-4o https://t.co/xhqFwqlAPV https://t.co/uasCCCsQRp