Will any AI model score above 90% on ARC-AGI-2 by end of 2026?

ARC-AGI-2 tests flexible reasoning beyond memorisation. Current frontier models plateau around 55–60 %. A breakthrough would signal a shift.

Yes28%72%No

Volume

$5.46M

Traders

31.5K

Liquidity

$1.2M

Yes Price

$0.28

Probability History

Select Outcome

Prices include spread.

Select an outcome to start trading

Amount (USD)

You Pay—

Fee (1%, incl.)—

Trade Amount—

Shares—

Current Price—

Max Payout—

Max profit if correct—

Mock mode — no real funds used

ARC-AGI-2 tests flexible reasoning beyond memorisation. Current frontier models plateau around 55–60 %. A breakthrough would signal a shift.

You can read comments as a guest, but posting requires an account

Loading comments…