Azernews.Az

by Alimat Aliyeva

The American company OpenAI has introduced a new version of its artificial intelligence model, GPT-4.5, Azernews reports.

The developer has called this release their "most knowledgeable model," though they note that GPT-4.5 is not a breakthrough and does not possess reasoning capabilities.

"We are releasing a preview version of GPT-4.5— our largest and most advanced chatbot model. GPT-4.5 improves its ability to recognize cues, make connections, and generate creative ideas without actual reasoning," the company stated.

This new version responds more effectively to user prompts, making it particularly useful for writing and programming tasks. OpenAI expects GPT-4.5 to make fewer mistakes compared to its predecessors.

Despite its limitations, GPT-4.5 significantly reduces hallucinations compared to GPT-4 and performs slightly better than the GPT-3.5 model. Additionally, the new version demonstrates more developed intuition and creativity. It also excels at understanding subtle cues and implicit expectations, offering responses that are more nuanced.

OpenAI emphasizes that GPT-4.5's improved ability to interpret subtle signals makes it an even more valuable tool for professionals in fields like content creation, software development, and data analysis. However, despite these improvements, OpenAI continues to emphasize the importance of human oversight, as the model can still occasionally produce inaccurate or biased information.