OpenAI recently announced that it will gradually launch its latest artificial intelligence model GPT-4.5 to ChatGPT Plus users. The news has attracted widespread attention, especially for technology enthusiasts and developers, the launch of GPT-4.5 marks another major advance in artificial intelligence technology. Sam Altman, CEO of OpenAI, said that due to model capacity and resource constraints, initial access will be open in phases to ensure the stability and quality of the user experience.
As the largest AI model of OpenAI to date, GPT-4.5 invests more computing resources and data during training, aiming to significantly improve its performance in conversational tasks. Compared with previous versions, GPT-4.5 demonstrates stronger abilities in daily chat, practical problem solving, and creative tasks. Additionally, OpenAI said the new model is expected to reduce the common "illusion" phenomenon in previous versions, namely the question of generating inaccurate or fictional answers.
However, despite significant progress in dialogue capabilities in GPT-4.5, OpenAI also admits that its performance in complex reasoning is still insufficient. Compared with competitors such as Anthropic's Claude and DeepSeek's AI models, GPT-4.5 seems to be incompetent in dealing with multi-step logic problems and complex reasoning tasks. "GPT-4.5 is not a reasoning model and cannot lead the benchmark," Altman made it clear on social media.
In addition to technical challenges, the pricing issue of GPT-4.5 has also sparked widespread discussion. It is reported that the model costs up to $150 per million generated tokens, which is 15 times more expensive than OpenAI's popular model GPT-4o. Such high costs have caused many users to doubt the possibility of their widespread adoption, especially given its limitations in logical reasoning.
Despite this, GPT-4.5 still has a unique appeal. OpenAI internal testing shows that the model performs particularly well in persuading other AI models, further highlighting its strong dialogue capabilities. In addition, GPT-4.5 fully supports ChatGPT tools and APIs, providing developers with great flexibility and innovation space.
The launch of GPT-4.5 is not only an exploration of OpenAI’s technology boundaries, but also reflects the company’s tradeoffs between strength, performance and real availability. ChatGPT Plus users will gain access in the next few days, and OpenAI will continue to evaluate its potential and challenges in technological development.
Key points:
GPT-4.5 is the largest artificial intelligence model launched by OpenAI and will be gradually opened to ChatGPT Plus users.
Although GPT-4.5 has significantly improved dialogue capabilities, there are still shortcomings in complex reasoning.
Using GPT-4.5 costs up to $150 per million tokens, raising concerns about its widespread adoption.