OpenAI released a new AI model, GPT-4o, with some interesting capabilities. It also maintains the OpenAI tradition of terrible names for AI models (the “o” means “omni” - more on that shortly). Previously, new AI models from the major AI labs have focused on how smart the model is. GPT-4o appears to be a step up over GPT-4 and is the smartest model I have used. However, it does not represent a major leap over the previous version of GPT-4, the way that GPT-4 was a 10x improvement over the free GPT-3.5. That has to wait, presumably, until GPT-5, which is apparently still scheduled for some future release. But what it does do is quite interesting.
No comments:
Post a Comment