In the displayed ranking, that ties it with Anthropic’s Claude Fable 5.1. Claude Opus 5.5 and OpenAI’s GPT-6 Astra share the top position.
The ECI combines results from many AI benchmarks into a single general-capability scale. Epoch says the index is intended to permit comparison across models and over periods in which individual benchmarks may become saturated. Its underlying benchmark set spans areas including agents, computer use, mathematics, software engineering, science, long context, and multimodal tasks.
The score is an abstract composite measure, not a result from one benchmark. Epoch’s page says ECI values should be interpreted through comparisons and expected benchmark-performance curves rather than in isolation.
The index page does not establish pricing, API availability, rate limits, deployment terms, or performance in a particular production workflow. It also does not by itself establish that a score difference will translate into a measurable advantage for a given user task.