GPT-4.5 AI Model Expands to More ChatGPT Users

OpenAI has kicked off the rollout of its latest AI model, GPT-4.5, to users on the ChatGPT Plus plan. In a bunch of posts on X, they mentioned that the rollout would take around "1-3 days," and they're planning to tweak the rate limits as they go. Last week, folks with the $200-a-month ChatGPT Pro subscription got first dibs on GPT-4.5.

"We're aiming to give everyone with access to GPT-4.5 a pretty generous rate limit, but we'll need to adjust those limits as we figure out how much demand there is," OpenAI shared in one of their posts.

Now, GPT-4.5 is the biggest AI model OpenAI has ever put out, using heaps more computing power and data than their past models. But here's the thing—it's not the top dog when it comes to performance. On some AI benchmarks, GPT-4.5 doesn't quite stack up against newer "reasoning" models from the likes of DeepSeek, Anthropic, and even OpenAI's own newer stuff.

Running GPT-4.5 is also pretty pricey, according to OpenAI. They're actually mulling over whether it's worth keeping it in their API long-term because of the costs. Right now, they're charging $75 for every million tokens (that's about 750,000 words) you feed into the model, and $150 for every million tokens the model spits out. That's 30 times more for input and 15 times more for output compared to their go-to model, GPT-4o.

Despite the costs, OpenAI says that GPT-4.5's bigger size has given it a "deeper world knowledge" and "higher emotional intelligence." Plus, it seems to hallucinate less than most models out there, which means it's less likely to just make things up.

And get this—GPT-4.5 is pretty slick when it comes to persuasion. One of OpenAI's own tests showed that the model is really good at sweet-talking another AI into handing over cash and spilling a secret code word.