OpenAI has announced "GPT-6 Astra," a next-generation model representing a significant leap in the company's AI capabilities. According to the company, GPT-6 Astra is the most intelligent model in the world, possessing state-of-the-art performance in areas such as computer use, browsing, software engineering, cybersecurity, science, and professional tasks.

Additionally, as part of its computing strategy, OpenAI is developing its first custom inference chip, "Jalapeño." It is reported that Jalapeño has increased peak token throughput per watt by 1.5 to 1.9 times and reduced end-to-end latency by 1.7 to 3.6 times compared to existing commercial systems. The company plans to begin the deployment of Jalapeño within the year, alongside accelerators from partner companies such as NVIDIA and AMD.

As a result of its research and development, OpenAI also revealed that one of its internal models produced a solution regarding the "Navier-Stokes equations," one of the unsolved problems in mathematics. Through improvements in model capabilities and the integrated optimization of software and hardware, OpenAI aims to provide AI services more cost-effectively and efficiently.

Source: