OpenAI has introduced GPT-4.5, its largest and most powerful artificial intelligence model to date. This new model has been trained with significantly more processing power and a larger data set than previous versions. Initially, GPT-4.5 is available to ChatGPT Pro users and developers subscribed to OpenAI's API service. However, in the near future, ChatGPT Plus and ChatGPT Team users will also be able to access this new model.
OpenAI highlights the vast amount of data and advanced structures used in the training of GPT-4.5. As a result, GPT-4.5 has a deeper and broader understanding of the world and demonstrates higher emotional intelligence. It provides users with responses in a more natural and sincere tone, and it also excels in creative tasks such as writing and design. One of its key strengths is its ability to better understand human intent, distinguishing it from other artificial intelligence models.
However, OpenAI acknowledges that running this model requires significant costs. Therefore, they are evaluating whether to continue offering GPT-4.5 through their API in the future. GPT-4.5 currently supports features like image and file uploads, though it does not yet support capabilities such as realistic two-way voice communication.
Sam Altman, CEO of OpenAI, emphasized that GPT-4.5 represents a major technological breakthrough but also has “a kind of magic we haven’t felt before.” The new model has outperformed GPT-4.0 and many other models in terms of accuracy in SimpleQA tests. GPT-4.5 experiences fewer hallucinations, meaning the likelihood of providing incorrect information is reduced.
OpenAI refers to GPT-4.5 as a "research preview," aiming to better understand its strengths and weaknesses. They also look forward to learning how the model will be used and how people will apply it in unexpected ways. GPT-5 is planned for release later this year, which is expected to bring even greater advancements in artificial intelligence.
