U.S. Government Allows Anthropic to Re-release Mythos 5 Model to a Limited Set of ~100 Institutions
English summary
The U.S. government, through Commerce Secretary Howard Lutnick, has authorized Anthropic to re-release its new model Mythos 5, but only to a limited group. Access will be granted to approximately 100 institutions, including government agencies and private companies. This decision follows a similar phased-release requirement imposed on OpenAI for its GPT-5.6 model, indicating a coordinated regulatory approach to advanced AI systems.
Chinese summary
美国商务部长Howard Lutnick致信Anthropic，批准其重新发布Mythos 5模型，但现阶段仅限少数客户。该模型将向约100家机构开放，其中包括政府机构和私营公司。此前OpenAI的GPT-5.6模型也面临类似的分阶段发布要求，显示出对先进AI系统采取协同监管的趋势。
Key points
Commerce Secretary Howard Lutnick authorized Anthropic to re-release Mythos 5 under restrictions.
商务部长Howard Lutnick批准Anthropic重新发布Mythos 5，但有限制条件。
Access is limited to about 100 institutions, including both government agencies and private companies.
访问权限仅限于约100家机构，涵盖政府机构和私营公司。
OpenAI's GPT-5.6 follows the same phased-release model, reflecting a pattern of U.S. government oversight on advanced AI models.
OpenAI的GPT-5.6也采用相同的分阶段发布模式，反映出美国政府对先进AI模型的监管举措。