According to the article, GLM-5.2 is a Chinese open-source model causing concern in the West. Artificial Analysis shows it uses 43,000 output tokens per task, of which 37,000 are for reasoning, more than GLM-5.1 (26,000), MiniMax-M3 (24,000), Kimi K2.6 (35,000), and DeepSeek V4 Pro max (37,000). An increasing number of Western companies use Chinese models for less complex tasks due to cost savings. On 12 June 2026, Anthropic announced the US government issued an export control directive to suspend access to Fable 5 and Mythos 5 for any foreign national on national security grounds, after hostile state actors (Chinese and Russian) used those models to identify critical vulnerabilities. The open weights approach of Chinese models allows local download and management, reducing risk of unilateral access revocation, but does not eliminate security, compliance, and geopolitical issues. For Western companies, using a Chinese model via API to providers like DeepSeek or Z.ai may expose texts, prompts, documents, or code to non-Western rules, contracts, and jurisdictions, raising issues of data protection, trade secrets, compliance, internal controls, and liability.