As “general intelligence” remains out of reach, OpenAI is leaning more on application-specific pitches. Health advice is its most dangerous test case.

GPT-5 is not significantly ahead of Chinese AI models, according to Zhang Linfeng, an AI professor at Shanghai Jiao Tong University.

As “general intelligence” remains out of reach, OpenAI is leaning more on application-specific pitches. Health advice is its most dangerous test case.

Since its debut, GPT-5 has more than doubled coding and agent-building activity and driven an eight-fold jump in reasoning workloads.