SimpleBench: leaderboard
Evaluates 'linguistic adversarial robustness' - ability to answer trick questions. 200+ questions where everyday human reasoning (83.7%) still outperforms frontier LLMs on spatio-temporal reasoning and social intelligence.
Metric: Score (AVG@5). Source: simple-bench.com. Status: saturation imminent. 99 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5.1 | 86.6 |
| 2 | Median Human | 83.7 |
| 3 | Gemini 3.8 Flash | 82.4 |
| 4 | Claude Fable 5 | 81.9 |
| 5 | Muse Spark 1.3 | 81.8 |
| 6 | Claude Opus 5 | 80.6 |
| 7 | Gemini 3.1 Pro (Preview) | 79.6 |
| 8 | GPT-5.5 Pro | 76.9 |
| 9 | Gemini 3.5 Flash | 76.7 |
| 10 | Gemini 3 Pro (Preview) | 76.4 |
| 11 | Grok 4.6 | 75.9 |
| 12 | Muse Spark 1.2 | 74.5 |
| 13 | Qwen 3.7 Max | 70.4 |
| 14 | Grok 4.5 | 70 |
| 15 | GPT-5.5 | 69 |
Interactive version: theaggregate.ai/benchmark?slug=simplebench · How It Works · Data refreshed daily, snapshot 2026-09-05.