METR Finds Anthropic's AI R&D Accelerating 1.9x, Sparking Reassessment and Redaction Controversy
Around November 6, METR released an independent assessment of whether Anthropic's AI is accelerating its own AI R&D, concluding that acceleration is indeed happening — and turning "recursive self-improvement" from a theoretical concept into a measurable one for the first time. The report triggered dramatic expectation updates across the AI community, while Anthropic's heavy redactions sparked a transparency controversy.
Confirmed
- Starting in August, the METR team spent roughly 11 weeks reviewing Anthropic's internal R&D data and interviewing core researchers, focusing on three questions: whether acceleration is occurring, whether it is attributable to AI itself, and whether the risk narrative holds up (m13, m14, m19).
- Key quantitative finding: as of October 2026, Anthropic's model capability growth rate was about 1.9× that of the March 2024–October 2025 period (95% CI 1.5×–2.8×), with the AECI trend line rising from +15 points per year to +29 points — equivalent to compressing roughly two years of historical progress into one (m3, m11, m20).
- METR rarely disclosed internal model data not yet released by Anthropic: the latest model, labeled Model 4, has been available internally since October 2026 with an AECI score of 178.6 (m5, m8).
- Key takeaways relayed by rohanpaulai indicate that about 80% of the acceleration comes from AI itself (m14).
- Informal surveys and Slack polls compiled by METR show Anthropic employees self-reporting efficiency gains from Model 4 growing from 3.7× in early 2025 to around 12×, though the authors stress this is only very loose evidence (m17).
- METR said it submitted a roughly 100-page full report to Anthropic, but most of what it considered the most informative evidence was cut from the public version; Anthropic said the redactions removed confidential information that could "accelerate other AI developers" (m9). The R&D section of Anthropic's August 2026 risk report was also heavily blacked out (m1).
- Anthropic's co-founder responded to the redaction controversy, acknowledging it makes METR's work harder for others to verify, but said disclosing "the drivers of the acceleration" would be inappropriate given commercial sensitivity and safety concerns (m6, m18).
Community Reaction and Controversy
- Andrew Curran observed that many in the AI world made drastic updates to their RSI/ASI timelines because of the report, likening it to the shock of ChatGPT's release in November 2022; he said several former skeptics now believe RSI is less than 6 months away (m2, m4). He then warned the safety camp that safety plans relying on friendly governments, international agreements, or time windows beyond a year no longer make sense (m7).
- Investor Jake Mendel posted a long "mea culpa," saying he had previously been more pessimistic about timelines than his friends inside Anthropic and that his stance is now shaken; Gary Marcus questioned claims of "superintelligence within 18 months" (m10).
- A view shared by scaling01 argues that real-world progress is tracking the AI 2027 report's projections fairly closely (that scenario posits AI delivering roughly 2× R&D acceleration around December 2026), and the poster expressed unease about the pace (m15).
- New York State Assemblymember Alex Bores cited the report in calling on Congress to ban recursive self-improvement this year, noting OpenAI has said it expects to achieve recursive self-improvement by March 2028 (m16).
- Gary Marcus criticized Anthropic's co-founder for defending the redactions on safety grounds; among reader feedback relayed in m12, many said they felt "unsettled" after reading the report.
Why It Matters
This is the first time an independent evaluator has used a frontier lab's internal data to quantify "AI accelerating AI R&D" into a measurable conclusion, with data points (the 1.9× acceleration, the internal Model 4) drawn from models not yet public. Whether or not the findings ultimately hold, they have already materially shifted how researchers, investors, and even legislators judge AGI/ASI timelines — and exposed the structural tension between safety-evaluation transparency and trade secrets: the evaluator gets to see the internal data, while the public only sees the redacted version.
2026-11-06 ~ 2026-11-07 · 117 related posts
- Episode 1: Anthropic Says It May Imminently Hit RSP Threshold for AI-Accelerated R&D(2026-11-03, 7 posts)
- Episode 2: AI 2027 Authors Warn Anthropic's 2x Threshold Could Accelerate Superintelligence Timeline(2026-11-05, 2 posts)
- Episode 3: METR Finds Anthropic's AI R&D Accelerating 1.9x, Sparking Reassessment and Redaction Controversy(2026-11-06, 117 posts)
- Episode 4: Report: Anthropic's Next Model Finished Training, Tops All METR Benchmarks(2026-11-06, 2 posts)
- Episode 5: METR Says Anthropic Canceled Compute Allocation Assessment Pilot, Sparking Criticism(2026-11-06, 4 posts)
- Episode 6: AI Safety Community Criticizes Anthropic's Acceleration Strategy(2026-11-06, 2 posts)
- Episode 7: Oxford's Toby Ord Argues METR's 2x RSI Speedup Drops to 1.4x Once Pretraining Jumps Are Counted(2026-11-06, 11 posts)
- Episode 8: Oxford philosopher explains why AI benchmark error bars have widened(2026-11-06, 2 posts)
Primary sources
- [source] METR launches AI R&D acceleration assessment: are capabilities outrunning historical trends? — METR_Evals · 2026-11-06
- Anthropic shared extensive nonpublic AI R&D data with METR, incl. algorithmic progress estimates — METR_Evals · 2026-11-06
- METR: capability acceleration not driven by faster training compute growth — METR_Evals · 2026-11-06
- METR publishes risk assessment on whether AI is accelerating AI R&D — RyanGreenblatt · 2026-11-06
- METR publishes AI-accelerating-AI risk assessment; Anthropic redacted most key evidence — Miles_Brundage · 2026-11-06
- METR says Anthropic redacted most key evidence from its 100-page safety assessment report — tszzl · 2026-11-06
- METR: Anthropic models' capability growth hits 1.9x the earlier rate — dfrsrchtwts · 2026-11-06
- METR finds superexponential fit as AI drives current R&D acceleration, staff self-report gains — ricklamers · 2026-11-06
- ~15 staff interviews key to external evals of unreleased models' RSI capabilities — dfrsrchtwts · 2026-11-06
- METR chart reveals unreleased Anthropic Model 4 scoring 178.6 on AECI — ResultBackground2450 · 2026-11-06
- METR estimates AI is already accelerating AI development by roughly 2x — harris_edouard · 2026-11-06
- Anthropic Cofounder Sam McCandlish Responds to METR Report on AI Accelerating AI Development — edwinarbus · 2026-11-06
- METR: Anthropic likely hit its 2x RSP acceleration threshold, maybe 3x, ~80% AI-driven — daveholtz · 2026-11-06
- Rumor: Anthropic sitting on internal models that top every benchmark with 12x self-reported speedup — teortaxesTex · 2026-11-06
- METR finds Anthropic packing nearly 2 years of AI progress into 1 year, no clear bottlenecks — 141_1337 · 2026-11-06
- METR report reveals Anthropic's internal Model 4: trained Oct 2026, scores 178.6 on ECI — legit_api · 2026-11-06
- METR chart shows AI benchmark progress nearly doubled, Anthropic Model 4 hits 178.6 AECI — scaling01 · 2026-11-06
- METR says Anthropic redacted most of the key evidence from its safety assessment — YashasGunderia · 2026-11-06
- METR chart shows unreleased Anthropic "Model 4" at ~12x self-reported coding speedup — YashasGunderia · 2026-11-06
- METR: Anthropic redacted most key evidence from public version of 100-page safety audit — Miles_Brundage · 2026-11-06
- METR report: AI is accelerating AI capability gains inside Anthropic — mattbeane · 2026-11-06
- METR report: RSI already accelerating AI progress ~2x, matching AI 2027's March 2027 mark — RyanGreenblatt · 2026-11-06
- METR: Anthropic's new internal Model 4 tops all benchmarks, may trigger 3x scaling threshold — scottleibrand · 2026-11-06
- Miles Brundage cites METR: Anthropic will pack 4 months of progress into 2 before Congress — nitarshan · 2026-11-06
- 'Recursive Self-Improvement Likely Began October 2025,' Citing METR Data — ben_j_todd · 2026-11-06
- METR Backs Anthropic's ~2x R&D Acceleration Claim, Sees ~80% AI-Driven, ASI by 2028 — ben_j_todd · 2026-11-06
- Report: AI is doubling Anthropic's capabilities progress, a big update toward very short timelines — EvanHub · 2026-11-06
- METR's Greenblatt: AI R&D acceleration already at AI 2027's March 2027 levels — ResultBackground2450 · 2026-11-06
- METR: Anthropic's model capability gains have ~2x'd, no bottlenecks observed — corbtt · 2026-11-06
- METR: majority of capability acceleration likely driven by AI doing AI R&D — nathanbenaich · 2026-11-06
- AI 2027 author: METR data shows progress even faster than forecast — DKokotajlo · 2026-11-06
- METR report on AI accelerating AI R&D flips a skeptic: >50% on superintelligence in 18 months — GaryMarcus · 2026-11-06
- METR report finds AI is accelerating AI R&D at Anthropic; researcher now puts >50% odds on superintelligence within 18 months — EvanHub · 2026-11-06
- METR report: Anthropic is speeding up, not pacing, and redacted the key parts — asusarla · 2026-11-06
- 7 of Anthropic's last 8 frontier models never got public evals — taoroalin · 2026-11-06
- 7 of Anthropic's last 8 frontier models skipped public evals, notes researcher — jessi_cata · 2026-11-06
- METR-style fit shows benchmark gains accelerating 1.9x, but sensitive to measurement — jessi_cata · 2026-11-06
- METR chart: Anthropic employees self-report ~12x productivity uplift from Model 4 — scaling01 · 2026-11-06
- METR chart: permanent acceleration fits Anthropic's capability curve better (RMSE 1.9 vs 3.9) — scaling01 · 2026-11-06
- METR: AI capability acceleration not driven by faster training compute growth — BenBajarin · 2026-11-06
- METR chart shows Anthropic self-reported uplift hitting 12x with mystery 'Model 4' — maksym_andr · 2026-11-06
- METR's first deep assessment of AI accelerating AI R&D interviewed 15 staff — but skipped a key question — ronbodkin · 2026-11-06
- [source] Anthropic cofounder defends METR report redactions; Gary Marcus unimpressed — GaryMarcus · 2026-11-06
- METR's Anthropic assessment triggers community-wide AGI timeline updates, veteran compares it to ChatGPT moment — AndrewCurran_ · 2026-11-06
6 near-duplicate retellings: Recoil42 · kimmonismus · scottleibrand · scaling01 · maksym_andr · RishiBommasani