GPT-4.5 Launch – OpenAI Reveals Latest AI Model for Enhanced Performance

GPT-4.5 has been officially announced by OpenAI as a research preview for ChatGPT Pro users and developers globally, with plans to broaden access in the upcoming weeks. The Microsoft-supported AI company confirmed this launch on Thursday, highlighting enhancements in pattern recognition, creative insight generation, and emotional intelligence.

Access Plan for GPT-4.5

Currently, GPT-4.5 is available to ChatGPT Pro users. OpenAI intends to provide access to ChatGPT Plus and Team subscribers next week, followed by Enterprise and Edu users the subsequent week.

Insights from the CEO on GPT-4.5

CEO Sam Altman referred to GPT-4.5 as “a giant, expensive model,” indicating that the company encountered GPU shortages, which affected its capability to launch the model for all user tiers at once.

Altman stated, “We will add tens of thousands of GPUs next week… this isn’t how we want to operate, but it’s hard to perfectly predict growth surges that lead to GPU shortages,” in a post on X (formerly known as Twitter).

Improvements in Hallucination Rate

A notable advancement in GPT-4.5 is its lowered hallucination rate, a common issue for large language models where they may produce inaccurate or misleading information. OpenAI claims that the hallucination rate for GPT-4.5 is now 37.1%, a significant improvement compared to the 61.8% rate recorded by GPT-4 and 44% by the o1 reasoning model.

Features of GPT-4.5

The GPT-4.5 model allows for file and image uploads and can assist in writing and coding tasks. However, features involving voice and video, which OpenAI is investigating, are not included in this release.

Future Expansion Plans

As broader access is anticipated in the coming weeks, OpenAI is preparing for a significant GPU expansion to guarantee consistent availability for its users. GPT-4.5 is expected to deliver enhanced performance across various domains, marking a vital step towards the future release of GPT-5.