Noometry

Reasoning benchmark

EnigmaEval leaderboard

As of October 2026, Claude Fable 5 has the highest published EnigmaEval score on Noometry at 39.3%, out of 38 models with results.

Last verified

About EnigmaEval

Long multimodal puzzle-hunt problems that require creative, multi-step reasoning.

Category
Reasoning
Introduced
2025
Format
Puzzles
Unit
Percent (random guessing ≈ 0%)
Official site
scale.com

Top 15 models

  1. Claude Fable 5 39.3%
  2. GPT-5.6 Sol 37.1%
  3. Gemini 3.1 Pro Preview 36.8%
  4. Gemini 3.5 Flash 25.4%
  5. GPT-5.4 Pro 23.8%
  6. Claude Opus 4.8 23.5%
  7. GPT-5 Pro 18.8%
  8. Gemini 3 Pro 18.2%
  9. GPT-5.4 16%
  10. o3 13.1%
  11. Claude Opus 4.5 11.9%
  12. GPT-5.1 11.2%
  13. GPT-5 10.5%
  14. GPT-5.2 10.4%
  15. o4-mini 9.2%

Sponsored placements are available on pages like this one. Advertise on Noometry

All results

#ModelProviderScoreSettingSourceDate
1Claude Fable 5 Anthropic39.3%highEpoch AI
2GPT-5.6 Sol OpenAI37.1%highEpoch AI
3Gemini 3.1 Pro Preview Google36.8%Epoch AI
4Gemini 3.5 Flash Google25.4%highEpoch AI
5GPT-5.4 Pro OpenAI23.8%Epoch AI
6Claude Opus 4.8 Anthropic23.5%xhighEpoch AI
7GPT-5 Pro OpenAI18.8%Epoch AI
8Gemini 3 Pro Google18.2%Epoch AI
9GPT-5.4 OpenAI16%xhighEpoch AI
10o3 OpenAI13.1%mediumEpoch AI
11Claude Opus 4.5 Anthropic11.9%Epoch AI
12GPT-5.1 OpenAI11.2%Epoch AI
13GPT-5 OpenAI10.5%Epoch AI
14GPT-5.2 OpenAI10.4%Epoch AI
15o4-mini OpenAI9.2%highEpoch AI
16GPT-5 Mini OpenAI8.2%Epoch AI
17Claude Opus 4.6 Anthropic7.6%maxEpoch AI
18Claude Opus 4.1 Anthropic7.2%Epoch AI
19o1-pro OpenAI6.1%Epoch AI
20Claude Sonnet 4.5 Anthropic6%Epoch AI
21o1 OpenAI5.7%Epoch AI
22Claude Opus 4 Anthropic5.6%Epoch AI
23Gemini 2.5 Pro Google5.6%Epoch AI
24Claude 3.7 Sonnet Anthropic4.2%Epoch AI
25Kimi K2.5 Moonshot AI3.4%Epoch AI
26GPT-4.5 OpenAI3.2%Epoch AI
27Claude Sonnet 4 Anthropic3.1%Epoch AI
28Gemini 3.1 Flash Lite Google3%Epoch AI
29Gemini 2.5 Flash Google2.7%Epoch AI
30GPT-4.1 OpenAI2.2%Epoch AI
31Gemini 2.0 Flash (Feb 2025) Google1.1%Epoch AI
32Claude 3.5 Sonnet Anthropic0.9%Epoch AI
33Pixtral Large Mistral AI0.8%Epoch AI
34Claude 3 Opus Anthropic0.8%Epoch AI
35GPT-4o OpenAI0.8%Epoch AI
36Gemini 2.0 Pro Google0.7%Epoch AI
37Llama 4 Maverick Meta0.6%Epoch AI
38Llama 3.2 90B Meta0.4%Epoch AI

Compare the leaders

Other reasoning benchmarks

Frequently asked questions

What does EnigmaEval measure?

Long multimodal puzzle-hunt problems that require creative, multi-step reasoning.

Which model has the highest EnigmaEval score?

As of October 2026, Claude Fable 5 has the highest published EnigmaEval score on Noometry at 39.3%, out of 38 models with results.

What is the best open-weight model on EnigmaEval?

Kimi K2.5 has the highest EnigmaEval accuracy among open-weight models at 3.4%, ranking 25 of 38 overall.