Open AI’s GPT-4.5 is Here for Pro Users

GPT-4.5 will start rolling out to ChatGPT Plus members “over a few days” as opposed to all at once, according to OpenAI CEO Sam Altman.

According to reports, GPT-4.5, the latest release of OpenAI’s large language model (LLM), will be the final one to skip chain-of-thought prompting. Chain-of-thought models break down difficult problems into manageable, logical stages that are followed to arrive at a solution in an effort to mimic human reasoning.

Even though GPT-4.5 is still in its infancy, OpenAI claims that preliminary testing shows that using it “feels more natural.”

“Its broader knowledge base, improved ability to follow user intent, and greater ‘EQ’ make it useful for tasks like improving writing, programming, and solving practical problems,” OpenAI stated last week. 

There is no surprise in the rolling, staggered release of GPT-4.5. According to Altman, OpenAI had originally intended to simultaneously introduce the “giant, expensive model” to ChatGPT Plus and Pro members, but this was not possible due to a shortage of graphics processing units (GPUs). As a result, ChatGPT Pro users were the first to obtain it. A monthly membership to ChatGPT Pro costs $200, but a monthly subscription to ChatGPT Plus costs $20.

The advantages of utilising GPT-4.5 are not immediately apparent when contrasting AI benchmark tests from rival models with OpenAI’s “reasoning” versions. However, according to the OpenAI researchers that worked on GPT-4.5, the model’s system card shows advancements in the anthropomorphic features, such as a richer comprehension of emotion and a stronger intuition. This was not the 2025 release anticipated after the company’s strategy to further commercialise ChatGPT as practical software. This style aims for a more emotive approach instead of a more practical one.

Given that OpenAI has been progressively increasing the number of business contracts it has, it is reasonable to anticipate that the firm would deliver flagship products with useful, built-in applications, particularly inside the most costly and potent version of its chatbot. GPT-4.5, on the other hand, is more in line with the results of an academic research team that is putting all of their effort into pursuing artificial general intelligence—a theoretical version of the algorithm that is proficient enough to take the place of white-collar workers and nearly divine in its capacity to process data.   

Out of all the intangible gains, GPT-4.5 is much better at creating an attention-grabbing headline. GPT-4o produced outputs that were less varied and less engaging overall.

OpenAI is searching for new methods to recover the cost of training and operating increasingly expensive frontier models as it develops GPT-5, which is expected to be a more unified model that combines “reasoning” with other research advancements.