Will the highest Epoch Capabilities Index (ECI) score in Epoch AI's benchmark data reach at least 170.0 for a model released on or before 2026-12-06?
échéance 2026-12-06
confiance moyenne
preuve niveau 4
51 %estimation PolySignal
Critère de résolution
Resolves YES if, in a snapshot of Epoch AI's benchmark data (https://epoch.ai/data/benchmark_data.zip, file "epoch_capabilities_index/eci_scores.csv", column "eci") taken on or after 2026-12-20, at least one model with a release date on or before 2026-12-06 has a value >= 170.0. Reference at creation: 167.3 (Claude Opus 5.5, 2026-09-22). Voided if Epoch AI revises the methodology so that values are no longer comparable, or retires the file. Data: Epoch AI, CC BY 4.0.
État de départ vérifié
Computed base rate: 45% of past 60-day windows in Epoch AI's benchmark data saw the running record rise by at least the required amount (threshold 170, current record 167.33). Data: Epoch AI, CC BY 4.0, epoch.ai/benchmarks. Overlapping windows: small effective sample.
Les arguments ci-dessous ont été produits par le modèle en anglais. Ils sont affichés tels quels plutôt que traduits automatiquement, pour que rien sur cette page ne soit un texte que PolySignal n'a pas réellement écrit.
Ce qui fait monter
- The required increase from the confirmed baseline of 167.33 to 170.0 (+2.67 ECI points) is relatively small and achievable within a 60-day window during active release cycles.
- Major frontier labs frequently ship updated or fine-tuned model iterations towards the end of the calendar year.
Ce qui fait descendre
- Claims of newer models achieving scores near 169 (such as Astra [2]) come solely from tier 4 media and lack tier 1 or tier 2 corroboration.
- Frontier model release schedules are subject to delays, and 60 days provides a narrow window for a confirmed release with >=170 ECI.
Ce qui ferait le plus bouger ce chiffre
Whether major frontier AI developers (e.g. Anthropic, OpenAI, Google) release a major model update before December 6, 2026 that is evaluated on ECI.
Comment ce chiffre est fabriqué
| Échéance | 2026-12-06 |
| Taux de base | 45% |
| Meilleure preuve | niveau 4 · 1 article(s) utilisé(s) |
| Estimations | 50 · 52 (écart 2 pts) |
| Confiance | moyenne |
| Plafond de preuve | non déclenché |
| Révisée le | 2026-10-07 23:54 |
| Modèle | gemini-3.6-flash |
Sources consultées