operational metric

Anthropic R&D Automation Index

Model-rated automation across a fixed basket of R&D work.

Observed R&D automation

What this measure tells us

Describes actual reported organizational use.

Benchmark version

Metric

Results to compare

Protocol

2026-09-17 · 378 tasks

378 leaf categories; work basket built from July records.

Work rated AI leads (percent): A task-specific measurement; not a percentage of RSI achieved.

Work rated fully autonomous (percent): A task-specific measurement; not a percentage of RSI achieved.

Published results · August 2026 snapshot
Key	System / organization	Metric	Reported result	Date	Protocol	Verification	Evidence
—	AnthropicAnthropic model R&D work	Work rated AI leads	26%	2026-08	automation-ai_leads	lab reported	Measurements for understanding the pace of AI development inside frontier labs ↗
—	AnthropicAnthropic model R&D work	Work rated fully autonomous	No measured subset fully autonomous	2026-08	automation-autonomous	lab reported	Measurements for understanding the pace of AI development inside frontier labs ↗

No reference points are indexed for this measure.