Which organization's model will have the highest Epoch Capabilities Index (ECI) score on Epoch AI's ECI page on November 30, 2026?

Resolution source: Epoch AI's ECI page (https://epoch.ai/eci) and the ECI scores file inside Epoch's downloadable benchmark data (https://epoch.ai/data/benchmark_data.zip, file epoch_capabilities_index/eci_scores.csv, column eci). I will check both on December 1, 2026 (UTC), using the data as published for November 30. If they disagree, the CSV wins.

How it resolves: to the organization (the CSV's Organization column) of the single model with the highest ECI score. Any model counts — open or closed, any version — as long as Epoch lists it.

Ties: if the top two scores are exactly equal in the CSV, the market resolves 50/50 between those organizations.

Edge cases:

• Epoch sometimes re-estimates all ECI scores when new benchmark data comes in, so the order of older models can change. Only the numbers published for Nov 30 count.

• A model released in November only counts if Epoch has scored it by Nov 30.

• If Epoch takes the ECI page and data down or stops updating it before Nov 30, I will use the last published version and say so in the comments.

Where things stand (Epoch data downloaded Oct 31, 2026): GPT-6 Astra (OpenAI) 169.1, Claude Opus 5.5 (Anthropic) 168.4, GPT-6.1 Terra (OpenAI) 168.1, GPT-6.1 Sol (OpenAI) 167.8, Claude Sonnet 5.5 (Anthropic) 166.7, Gemini 4 Argon (Google DeepMind) 166.1. The top two are less than a point apart, and their confidence intervals overlap heavily.

I will not trade on this market.