METR: Anthropic AI R&D capabilities accelerating about 2× prior pace
In a high-access assessment covering 12 August–29 October 2026, METR says Anthropic appears to be improving well-specified-task capabilities roughly twice as fast as from March 2024 to October 2025, with AECI benchmark growth about 1.9× since October 2025.
METR published on 5 November 2026 a high-access assessment of Anthropic AI R&D acceleration covering 12 August–29 October 2026. Subjectively, METR says Anthropic is likely improving capabilities on well-specified tasks at about 2× the rate seen from March 2024 to October 2025 (METR would not be surprised by a range of about 1.3×–3×).
What METR measured
On its AECI benchmark, growth was about 1.9× since October 2025 (95% CI 1.5–2.8×). METR says the acceleration does not appear to come from faster training-compute scaling. It guesses that a majority of the current acceleration — roughly 80%, with high uncertainty — likely comes from AI-assisted R&D, still mostly assistance rather than fully autonomous research. METR sees no clear bottlenecks for the next 12 months, though that judgment is uncertain.
Limits
Anthropic redacted much of the most informative evidence in the roughly 100-page full report. METR notes the assessment may over- or underestimate acceleration and that gains may be limited to well-specified tasks. These are METR’s findings from privileged access, not a public open evaluation.
Also available in: Саха тылаРусскийEspañol中文Portuguêsالعربية