ARC Prize Verified

Grok 4.5

xAI·Jul 16, 2026·3 reasoning variants

As of July 2026, Grok 4.5 is the latest model from xAI. It performs competitively on ARC-AGI-1. Raising the reasoning effort from medium to high did not result in increased performance or cost.

ARC-AGI 3 leaderboard

Grok 4.5

Verified scores

VariantARC-AGI-1ARC-AGI-2ARC-AGI-3
High
85.7%
52.6%
0.30%
Medium
87.2%
52.6%
0.32%
Low
79.2%
33.1%
0.26%

Tasks & environments

Pass/fail per reasoning level across each benchmark. Hardest tasks (fewest levels solving) are listed first.

ARC-AGI-3 Public Demo

25 environments

ARC-AGI-2 Public Eval

120 tasks
Task
High
Medium
Low

ARC-AGI-1 Public Eval

400 tasks
Task
High
Medium
Low