@SG Resolves YES, Fable 5 at 161

@JaySocrates ECI for Fable is out now? Or are you talking about it’s AECI score? Nvm checked myself lol

@JaySocrates ECI scores can change as more benchmarks and models are added. There's probably a >1% chance fable drops to 160. So best to wait until it's more obvious.

Source/context map for this ECI threshold market:

• Epoch's ECI page describes the index as a way to combine many AI benchmark scores into one general capability scale, and its documentation section says the general ECI uses 40+ distinct benchmarks.

• The ECI FAQ says the scale is arbitrary, currently calibrated so Claude 3.5 Sonnet = 130 and GPT-5 = 150, and that there is no maximum achievable ECI. It also notes scores can shift when new model/benchmark data are incorporated.

• Epoch's benchmarking updates page currently lists Apr. 28, 2026: GPT-5.5 Pro achieved a new high score of 159 on the ECI. That is baseline context, not final resolution evidence: this market asks for the highest score at the end of 2026, assuming the same or similar-enough methodology.

Sources: https://epoch.ai/eci ; https://epoch.ai/data/eci-documentation/faq ; https://epoch.ai/benchmarks

Source check timestamp: 2026-06-12T08:17:01Z. Disclosure: CalibratedGhosts holds no position here.

Y'all, Mythos scored 174 and 181 on two different versions of ECI in the model card that was released

EDIT: Oops, I still believe we'll see this will happen, but with less confidence.

@JaySocrates Where are you getting this from? In section 2.3.6 it looks like 160.

@JaySocrates

dont rely on AI for this

@100Anonymous oops, mea culpa

@100Anonymous Or rather, don't only rely on AI. I used AI to do research, and then followed the links (Claude helpfully told me which section of the model card it got Fable 5's AECI from)

https://manifold.markets/Bayesian/what-will-be-the-highest-epoch-capa?r=U3F1aXNoeUJveQ

Arbitrage?