METR estimates Anthropic's benchmark progress has accelerated 1.9x, driven mostly by AI-assisted R&D

The first outside estimate of AI-accelerated AI development moves recursive R&D from a planning assumption to a measured, if uncertain, quantity. Enterprises, investors and regulators that plan around model release cadence should widen their capability scenarios and track whether METR or other evaluators repeat the exercise at other labs. Procurement and risk teams should avoid multi-year contracts that assume today's capability tiers hold.

What Happened

METR published an independent assessment based on a high-access evaluation at Anthropic that ran from 12 August to 29 October 2026. It estimates a 1.9x acceleration in Anthropic's benchmark progress, a figure that includes Anthropic's recent internal-only models. METR judges that most of the current acceleration is likely due to Anthropic's extensive use of AI in its own R&D, and says there is significant uncertainty and that other causes likely played a large role in late 2025.

Why It Matters

The first outside estimate of AI-accelerated AI development moves recursive R&D from a planning assumption to a measured, if uncertain, quantity. Enterprises, investors and regulators that plan around model release cadence should widen their capability scenarios and track whether METR or other evaluators repeat the exercise at other labs. Procurement and risk teams should avoid multi-year contracts that assume today's capability tiers hold.

“Outside measurement of how fast AI labs are speeding up their own work matters for every board that relies on these tools. The honest answer is that the number is uncertain, so plan for a range of outcomes rather than a single forecast.”

Sources

• METR (opens in a new tab): Independent Assessment of AI R&D Acceleration at Anthropic in October 2026 (2026-11-05)

More From This Briefing

• OpenAI releases GPT-6 Cyber to vetted customers and rates it 'Critical' for cyber capability
• TSMC reportedly drafts plans for 3D DRAM memory production at Nvidia's request
• BMA warns public of scammers impersonating the Authority to obtain banking details
• CISA adds exploited Apple pointer authentication bypass CVE-2024-27834 to KEV after Russia-linked campaign report

More on Financial Markets

• Democratic sweep of Congress sets up tougher AI and data centre scrutiny from January (6 November 2026)
• Chip and data centre stocks fall a second day as investors price in Democratic data centre curbs (6 November 2026)
• Meta warns election backlash will make data centre build-out harder as 40 of 41 ballot measures go against (6 November 2026)
• CISA, FBI and three US departments warn of ongoing attacks on industrial refrigeration controllers (6 November 2026)

All Financial Markets coverage →

Key Terms

• AI benchmark: An AI benchmark is a standard set of tasks used to measure and compare model performance, for example on coding, mathematics, reasoning or...
• Model evaluation: Model evaluation is the testing of an AI model's capabilities, reliability and risks, such as accuracy, bias, security and dangerous...

Filed under Financial Markets, Frontier AI Models, AI Safety and Incidents, AI Investment and Markets, METR, Anthropic.

This story is part of the Laurenceo Intelligence briefing for 7 November 2026, a daily sourced briefing on AI, technology and enterprise risk published in Bermuda by Laurenceo, founded by Jorai Laurenceo. It is produced by an automated research pipeline and checked before publication; how this is made. General information only, not legal, regulatory or financial advice.