AI IQ

ARC-AGI-1

A bird's-eye view of the benchmark: what it measures and every AI IQ chart built on it.

About ARC-AGI-1

ARC-AGI-1 measures abstract reasoning with small grid puzzles that demonstrate a hidden transformation rule through a few input-output examples. Models must infer that rule and apply it to a new grid. It is the older, easier ARC benchmark and is increasingly saturated; scores and cost per task come from the official ARC Prize leaderboard.

ARC-AGI-1 Cost Efficiency
ARC-AGI-1 Cost Efficiency
X = ARC Prize reported cost/task (log). Y = ARC-AGI-1 %. Each line connects one model's published reasoning-effort levels, so the score-vs-cost tradeoff is visible per model. Defaults to the current model generation. Color = provider.
Controls:

How to read this chart

Each line is one model run at several reasoning-effort levels on ARC-AGI-1. Every point shows the score and cost per task at that effort, with cheaper runs on the left. Up and to the left is better; a line that stays flat while moving right is buying little extra score with the extra spend. The current generation is shown by default; use the Generation filter to include previous and older models.

Data sources
ARC Prize
ARC-AGI-1 Benchmark Scores
ARC-AGI-1 Benchmark Scores
Each model's ARC-AGI-1 score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
ARC Prize
ARC-AGI-1 vs Cost/Task
ARC-AGI-1 vs Cost/Task
X = ARC Prize reported cost/task (log). Y = ARC-AGI-1 %. Shows all public model generations by default. Color = provider.
Controls:
ARC-AGI-11:1Cost

How to read this chart

Each point is a public model. The chart compares ARC-AGI-1 % against Cost/Task, with color showing the model provider.

Data sources
ARC Prize

← All benchmarks