OpenAI has expanded access to its newest AI model, GPT-4.5, for ChatGPT Plus subscribers. The model first became available to users on the $200-a-month ChatGPT Pro plan. Now, it’s gradually reaching a broader audience.
A Quick Rollout, But What’s New?
OpenAI confirmed the rollout will take 1 to 3 days. However, they also noted that rate limits might shift as they track demand. This means users could see adjustments in how much they can interact with GPT-4.5 over time.
GPT-4.5 is OpenAI’s most advanced model, trained with significantly more data and computing power than its predecessors. It offers a broader knowledge base and improved emotional intelligence. But while it’s a step forward in some areas, it doesn’t necessarily outshine all other AI models.
A Bigger AI Model, But Not the Best
Despite its size and sophistication, GPT-4.5 doesn’t dominate every performance benchmark. Competing AI models from companies like DeepSeek and Anthropic have shown stronger reasoning capabilities in certain tests.
This raises questions about whether OpenAI’s focus on scaling up truly leads to better results. Still, the company stands by GPT-4.5’s enhanced abilities, particularly in understanding nuanced information and responding with greater emotional depth.
The Cost of Cutting-Edge Technology
GPT-4.5 is powerful but also costly. OpenAI has acknowledged that running the model is expensive. The expense is so high, that they’re still deciding whether to keep it available in their API long-term.
To cover expenses, OpenAI is charging $75 per million tokens (roughly 750,000 words) for input and $150 per million tokens for output. These rates are much higher than those for GPT-4, with input costs 30 times higher and output costs 15 times higher.
What’s Better About GPT-4.5?
Even with its high cost, GPT-4.5 comes with notable improvements. It “hallucinates” less frequently, meaning it’s less likely to generate incorrect or misleading information.
It also excels in rhetorical tasks. OpenAI’s internal tests show that GPT-4.5 is particularly skilled at persuading other AI models to perform tasks, like revealing a secret code word. While this might sound like a novelty, it highlights the model’s advanced reasoning and negotiation skills.
GPT-4.5 and The Future
The future of GPT-4.5 is still uncertain. OpenAI hasn’t decided whether it will remain in their API due to the high operational costs.
For now, users can explore its enhanced capabilities, but there’s no guarantee it will be available indefinitely. OpenAI is carefully weighing performance, demand, and costs to determine the best path forward.
