ARC Prize Verified

Inkling Small

Thinking Machines·Jul 30, 2026·6 reasoning variants

As of July 30, 2026, Thinking Machines' Inkling Small is the highest-scoring open-weight model evaluated by ARC Prize on both ARC-AGI-1 and ARC-AGI-2, while advancing the cost-performance frontier. At xhigh effort, it scores 84.0% on ARC-AGI-1 at $0.11 per task and 40.1% on ARC-AGI-2 Semi-Private at $0.23 per task, surpassing the original Inkling's scores of 79.5% at $0.30 and 36.5% at $0.64, respectively. It has not yet been evaluated on ARC-AGI-3.

Related models: Inkling

ARC-AGI 2 leaderboard

Inkling Small

Verified scores

VariantARC-AGI-1ARC-AGI-2ARC-AGI-3
Extra High
84.0%
40.1%
High
78.0%
33.1%
Medium
67.0%
13.6%
Low
52.5%
4.9%
Minimal
48.5%
4.0%
None
5.5%
0.0%

Tasks & environments

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

ARC-AGI-2 Public Eval

120 tasks
Task
Extra High
High
Medium
Low
Minimal
None

ARC-AGI-1 Public Eval

400 tasks
Task
Extra High
High
Medium
Low
Minimal
None