Anthropic Publishes Three Internal Metrics to Track AI R&D Automation and Oversight

AnthropicAI · x · 2026-09-18

Anthropic is publishing three measurements to illuminate how AI is increasingly used to build its next version: the share of AI R&D done by AI, how well AI agents can be overseen, and how compute is allocated. It provides an internal snapshot of each metric, arguing any frontier lab could publish the same measures and third parties could verify them. Anthropic frames this as narrowing the gap between what frontier labs know and what the public knows, giving society a chance to decide how to pace the frontier. Full methodology is on its blog.

Related event: Anthropic Unveils First AI Self-Development Metrics: Claude Now Leads 26% of R&D(15 posts)→

Original post →

More from AGI Musings

AGI Musings channel →