Our position on open-weights models

Anthropic released a formal statement detailing its opposition to the widespread release of open-weights models, citing potential safety and security risks. The company argues that as models become more capable, their ability to assist in harmful activities like biological weapon production outweighs the benefits of transparency. This position sets Anthropic apart from industry peers who advocate for open-source development, signaling a continued commitment to a closed-model approach as the primary strategy for managing advanced artificial intelligence.