China is falling behind in the AI race, according to a US government benchmark

A U.S. government benchmark says DeepSeek V4 Pro trails the top U.S. AI models by roughly eight months.

The biggest performance gaps are in reasoning, cybersecurity, and coding, while math is closer.

The report says the U.S.-China AI gap is widening, though outside rankings show a more stable difference.

Price may become China’s key advantage if customers prioritize lower cost over peak capability.