Skip to content
AIMODELRACE
Hard Reasoning · sub-raceClear lead

Common-Sense Reasoning race

Trick questions about everyday physics, time and social situations that people find easy.

Leader
Fable 5
Margin
1.3 pts
1.6% ahead
Weeks at #1
20
in a row
Changes
2
in 2026

Warming up the track…

Oct 20, 2026Fable 5 leads
2025Apr ’25Jul ’25Oct ’252026Apr ’26Jul ’26Oct ’26

Score history

Top 6 today over the leader's frontier · click to jump to a week

#1 frontierFable 5Opus 53.1 Pro3.5 FlashGrok 4.6Spark 1.2
10.020.030.040.050.060.070.080.090.02025Apr ’25Jul ’25Oct ’252026Apr ’26Jul ’26Oct ’26

Reigns at #1

Longest: Fable 5, 20 weeks (from Jun 2026).

Sibling races

More hard reasoning races

Why this race matters

Catches models that are strong on exams but stumble on basic real-world reasoning.

How it is scored

SimpleBench average score via Epoch AI's Benchmarking Hub. As measured, placed at release date.

Unit: percentHigher is betterPhoto finish under 0.5 ptsDominant over 3.0 pts

Data provenance

SimpleBench · Epoch AI

As measured: one result per model, placed at its release date.

38 tracked models with a result (104 entries on the source board) · newest data August 12, 2026.

Sources, licences and credits