Anthropic ECI (AECI), November 2026 fit

Anthropic's ECI (AECI), Anthropic's fork of Epoch AI's Capabilities Index, fitted on Anthropic's internal benchmark results and published in its system cards. This is the November 6, 2026 fit from the Claude Fable 5.2 and Claude Mythos 5.2 system card (Table 2.3.5.3.A): 371 benchmarks, 10,803 observations and 1,026 models. The scale is fixed by Claude Sonnet 3.5 at 130 and GPT-5 at 150. Each refit re-estimates the whole scale, so values from earlier or later cards are not comparable and a new fit gets a new benchmark. Only the table's released models are recorded; Anthropic's unreleased internal research model is not. Kept outside the capability fit for external validation.