Reasoning benchmark
SimpleBench leaderboard
As of October 2026, Claude Fable 5 has the highest published SimpleBench score on Noometry at 81.9%, out of 77 models with results.
Last verified
About SimpleBench
Trick questions about space, time and social situations that most people answer easily but models often miss.
- Category
- Reasoning
- Introduced
- 2024
- Format
- Multiple choice
- Unit
- Percent (random guessing ≈ 16.7%)
- Official site
- simple-bench.com
Top 15 models
- Claude Fable 5 81.9%
- Claude Opus 5 80.6%
- Gemini 3.1 Pro Preview 79.6%
- GPT-5.5 Pro 76.9%
- Gemini 3.5 Flash 76.7%
- Gemini 3 Pro 76.4%
- Grok 4.6 75.9%
- Muse Spark 1.2 74.5%
- GPT-5.4 Pro 74.1%
- GPT-5.6 Sol 71.7%
- Qwen3.7 Max 70.4%
- Grok 4.5 70%
- GPT-5.5 69%
- Claude Opus 4.6 67.6%
- Claude Opus 4.8 64.8%
Sponsored placements are available on pages like this one. Advertise on Noometry
All results
Google
Meta
Alibaba (Qwen)
DeepSeek
MiniMax
Mistral AI
CohereCompare the leaders
Frequently asked questions
What does SimpleBench measure?
Trick questions about space, time and social situations that most people answer easily but models often miss.
Which model has the highest SimpleBench score?
As of October 2026, Claude Fable 5 has the highest published SimpleBench score on Noometry at 81.9%, out of 77 models with results.
What is the best open-weight model on SimpleBench?
DeepSeek V4 Flash has the highest SimpleBench accuracy among open-weight models at 61.1%, ranking 21 of 77 overall.