We’ve seen some pretty powerful AI models from OpenAI, and it doesn’t look like the company is stopping anytime soon. While it made some significant launches during its Shipmas last year, it still has some powerful models up its sleeve. OpenAI officially launched GPT-4.5.
This shouldn’t come as a surprise to someone who follows the company via social media. OpenAI recently announced that it was gearing up to officially ship GPT-4.5 and GPT-5 at some point in the year. While GPT-4.5 is official, GPT-5 is still several months away. We don’t know what that model will have, but we know that it will push the AI industry forward.
OpenAI officially launches GPT-4.5
While this is a notably powerful model, OpenAI says that GPT-4.5 is not a frontier model. This means that it won’t show the best that the company can offer. Rather, it will be an iterative upgrade to what we have now.
Only people who have the $200/month ChatGPT Pro plan can use it, which is a pretty big bummer. Judging by what the company says, people not on that plan will miss out on a rather capable model. GPT-4.5 is OpenAI’s “most knowledgeable model yet,” and it brings improvements to writing capabilities and world knowledge. Also, the company states that the new model brings more personality over its predecessors. That sounds bittersweet, as it could be used to more closely emulate a human being.
While it’s a more powerful model, OpenAI didn’t release a complete version of GPT-4.5. It released a research preview of it. As such, it states that it’s not perfect, and that it might not perform as well as its o1 or o3 mini models. So, if you’re planning on using GPT-4.5, then you’ll need to know that it’s still baking.
As per OpenAI’s naming scheme, it looks like GPT-4.5 is a proper predecessor to GPT-4, which was released a while back. While it’s not a frontier model, it still brings some significant improvements.
2025-02-28 15:07:35