OpenAI has introduced a new AI model: what is known about it.

OpenAI has introduced a new version of the language model GPT-4.5

The company OpenAI has updated its language model GPT-4.5, designed to improve the quality of texts and programming code. It will be made available first to selected users, reports the company's official blog.

GPT-4.5 responds better to prompts, allowing it to more effectively create texts and develop software. This version is expected to contain fewer errors compared to previous models.

Currently, the previous version GPT codenamed Orion is available to a limited circle of developers and subscribers of ChatGPT Pro.

OpenAI Vice President Nick Ryder stated in his interview with Bloomberg that the new model will show the same productivity increase as the transition from GPT-3.5 to GPT-4 in 2023.

During the development of the Orion model, OpenAI faced certain difficulties. The model did not always successfully perform programming tasks it was not trained for. In addition, the company encountered problems in finding new high-quality data for training the AI, which are critical for the further development of the technology.

OpenAI plans to create its own chips for artificial intelligence

According to Reuters, OpenAI plans to complete the development of its own chips to support artificial intelligence technologies by 2025. This will reduce the company's dependence on chip supplier Nvidia. After the development is completed, OpenAI plans to manufacture chips at the Taiwan Semiconductor Manufacturing Co. factories. Mass production is scheduled to start in 2026.

On January 21, U.S. President Donald Trump announced the establishment of a company called Stargate, which will invest $500 billion in the development of artificial intelligence in the country. The project will involve companies OpenAI, SoftBank Group, and Oracle. According to Trump, this will create 100,000 new jobs.