HOT RADAR
llmHEAT 7.9°

AI CURATED HOT STORY · 11/6/2026

Anthropic's Three Internal Models Leaked, Top Model 4 Surpasses Opus 5.5 on Capability Index

1 reports archived1 independent sourcesupdated 11/6/2026, 23:21:56
Synthesis & Latest Updates
1 Sources Cross-Validated

Third-party evaluator METR disclosed three unreleased Anthropic internal models, Model 2/3/4, trained between July and October this year. The strongest, Model 4, scores 178.6 on Anthropic's AECI capability index, roughly 3.9 points above Opus 5.5 released six weeks earlier—equivalent to about three months of normal progress. Co-founder Sam McCandlish acknowledged AI is accelerating Anthropic's own R&D, while noting the public report omits substantial detail.

LATEST/Third-party evaluator METR disclosed three unreleased Anthropic internal models, Model 2/3/4, trained between July and October this year. The strongest, Model 4, scores 178.6 on Anthropic's AECI capability index, roughly 3.9 points above Opus 5.5 released six weeks earlier—equivalent to about three months of normal progress. Co-founder Sam McCandlish acknowledged AI is accelerating Anthropic's own R&D, while noting the public report omits substantial detail.

Hourly heat curve

19 fully observed hours
Now
7.9
Peak
11.811-07
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. 量子位官网T2·78 pts
    • Model 4 scores 178.6 on AECI, about 3.9 points above public Opus 5.5, equal to roughly three months of capability progress
    • METR estimates Anthropic's capability growth rate since October 2025 is about 1.9x the prior pace
    • Model 4 exceeds 50% on all 15 benchmarks in its AECI suite, versus several sub-50% results for Opus 4.6 nine months earlier
    • Anthropic confirms the public report was heavily redacted, citing commercial sensitivity and avoiding an acceleration roadmap