A recent evaluation by the Center for AI Standards and Innovation (CAISI) at the U.S. Department of Commerce’s National Institute of Standards and Technology (NIST) has revealed that AI models developed by China’s DeepSeek lag significantly behind their U.S. counterparts in performance, cost, security, and adoption. This assessment underscores the United States’ dominant position in the global AI landscape, highlighting both technical and strategic advantages.
U.S. AI Leadership Affirmed
The evaluation, part of President Donald Trump’s AI Action Plan, showcases the superiority of American AI systems. Commerce Secretary Howard Lutnick emphasized, “American AI dominates, with DeepSeek trailing far behind. Relying on foreign AI is not only technically inferior but also poses risks to security and innovation. The Department of Commerce is committed to maintaining U.S. leadership by setting standards and driving progress in AI.”
This report aligns with the AI Action Plan’s directive for CAISI to rigorously assess frontier AI models from the People’s Republic of China (PRC) and compare them with U.S. systems. The findings highlight America’s edge in fostering innovative, secure, and competitive AI technologies.
DeepSeek’s Shortcomings
CAISI’s analysis focused on three DeepSeek models—R1, R1-0528, and V3.1—and compared them against four U.S. models: OpenAI’s GPT-5, GPT-5-mini, gpt-oss, and Anthropic’s Opus 4. The evaluation spanned 19 benchmarks, including public state-of-the-art metrics and private benchmarks developed in collaboration with academic institutions and federal agencies.
The results revealed that DeepSeek’s models underperform in critical areas such as efficiency, cost-effectiveness, and security. Notably, issues related to censorship and vulnerabilities in DeepSeek’s model responses raise concerns for application developers, consumers, and U.S. national security. Despite these limitations, DeepSeek remains a prominent PRC developer, contributing to the growing global adoption of Chinese AI models.
Strategic Implications
The CAISI report warns that reliance on foreign AI systems like DeepSeek’s could expose users to risks, including potential malign influence and security vulnerabilities. These findings underscore the importance of prioritizing domestically developed AI to safeguard national interests and maintain technological sovereignty.
CAISI’s mandate, as outlined in the AI Action Plan, includes evaluating the capabilities of both U.S. and adversary AI systems, assessing their global adoption, and identifying potential threats. The center serves as a vital hub for industry collaboration, facilitating testing, research, and the development of best practices for commercial AI systems.
Driving Global AI Standards
As a cornerstone of NIST’s mission to advance U.S. leadership in AI, CAISI plays a pivotal role in shaping global standards. By fostering innovation and ensuring robust security measures, the center helps solidify America’s position at the forefront of the AI revolution.
This evaluation not only highlights the technical superiority of U.S. AI models but also reinforces the strategic importance of continued investment in domestic AI development. As global competition intensifies, CAISI’s work will remain critical in securing America’s technological edge.
About The Author
