OpenAI is releasing a research preview of what it calls its “largest and best” chat model to date, GPT‑4.5, which scales unsupervised learning in pre-training and post-training. As a result, the new chat model has the ability to recognize patterns, draw connections, and generate creative insights without having to draw on time and energy consuming “reasoning.” GPT‑4.5 is currently available to ChatGPT Pro subscribers ($200 per month) and developers subscribing to OpenAI’s API tier. ChatGPT Plus and ChatGPT Team customers are expected to gain access this week.

OpenAI CEO Sam Altman said the company “was forced to stagger the model’s rollout “because it ran out of GPUs,” according to TechCrunch, which reports that in a post on X, Altman described GPT-4.5 as “‘giant’ and ‘expensive,’ requiring ‘tens of thousands’ more GPUs before additional ChatGPT users can gain access.”

The enormous size could be part of the reason “GPT-4.5 is wildly expensive,” suggests TechCrunch, which details that “OpenAI is charging $75 per million tokens (~750,000 words) fed into the model and $150 per million tokens generated by the model” — or “30x the input cost and 15x the output cost of OpenAI’s workhorse GPT-4o model.”

Developed under the code name Orion, GPT-4.5 “is designed to be more general-purpose” than the company’s STEM-focused reasoning models, OpenAI says in a research paper.

“The industry has held its collective breath for Orion, which some consider to be a bellwether for the viability of traditional AI training approaches,” TechCrunch reports in a separate article, noting that “GPT-4.5 was developed using the same pre-training technique for unsupervised learning that dramatically increases the amount of computing power and data during that phase.”

The Verge notes GPT-4.5 “will have better writing capabilities, improved world knowledge, and what OpenAI calls a ‘refined personality over previous models.’”

In a blog post, OpenAI writes that “interacting with GPT‑4.5 feels more natural,” with “improved ability to follow user intent and greater ‘EQ’” (emotional intelligence) to improve its skill “solving practical problems,” adding that “we also expect it to hallucinate less.”