SimpleBench leaderboard: which LLMs score highest

SimpleBench (trick questions that test common sense) currently has scores for 64 models. The top score is 81.9% by Claude Fable 5 (max); the median model scores 43.3% and the lowest scores 10.7%.

Compare all benchmarks side by side on the live leaderboard →

Among the ten highest scorers, the cheapest listed API price belongs to Qwen3.7 Max at $1.25 per million input tokens (score 70.4%).

What SimpleBench measures

SimpleBench poses short questions with a trick or trap that most people catch using common sense. It is built to expose models that pattern-match instead of reason.

How to read these scores

Scores are percentages (higher is better): the share of questions or tasks the model completed correctly. Where a model is listed at several reasoning efforts, only its highest-effort row is shown. See the methodology for how rows are chosen.

Where the data comes from

Epoch AI Benchmarking Hub (CC BY 4.0). Scores were last captured 2026-10-05; the Captured column gives each row's date.

Related benchmarks

• WeirdML
• EnigmaEval
• Humanity's Last Exam
• METR time horizon
• OTIS Mock AIME
• Fiction.liveBench