Anthropic proposes new transparency framework for frontier AI models

AI safety company calls for public disclosure requirements targeting largest model developers with specific revenue and spending thresholds.