Atlas

Benchmarks

← All benchmarks

EnigmaEval

General QA · 2025-02-13

Score on puzzle-hunt reasoning tasks with multimodal clues.

Top models (higher is better)

ModelScore
GPT-6 Astra69.7
Gemini 4 Argon64.6
GPT-6.1 Sol59.5
Claude Opus 5.557.8
Claude Fable 5.144.2
Claude Opus 543.9
Gemini 3.8 Flash37.2
GPT-5.537.2
Muse Wonder36.1
Gemini 3.7 Flash33.9
GPT-5.427.6
Gemini 3.5 Flash25.4
Kimi K323.6
Opus 4.823.5
GPT-5.6 Terra23.1
GPT-5.6 Luna19.1
Gemini 3 Flash Preview18.3
Gemini 3 Pro Preview17.8
Grok 4.517.0
Muse Spark 1.116.6
Opus 4.715.6
Muse Spark Open14.7
GPT-5.214.5
Opus 4.512.4
o311.9
GPT-5.111.7
GPT-510.5
O4 Mini9.2
Sonnet 4.69.1
GPT-5 Mini8.2
Grok 47.8
Grok 4.207.7
Opus 4.67.6
GPT-5.4 Mini6.9
Grok 4.36.1
Sonnet 4.56.0
o1 (source unspecified)5.7
Gemini 2.5 Pro5.6
Kimi K2.65.5
Grok 4.1 Fast5.2
Loading Atlas data…