跳过主要内容
评测来源/Mystery Game Puzzles
参与排名推理

Mystery Game Puzzles

Epoch AI · Mystery Game Puzzles 的公开模型评测结果。

官方评测
来源说明

该来源测量模型在「推理」领域的具体交付表现,原始成绩会和统一能力读数分开展示。

限制: 不同版本、推理设置和测试集可能导致分数变化,请结合数据时间阅读。

使用方式

参与排名的来源会影响能力分;参考、交叉参考和观察来源仅用于对比与核验。

本来源的模型成绩

Claude Fable 5.20.991.02026-11-10
GPT-6 Astra0.8486.82026-11-05
GPT-6.1 Sol0.883.92026-11-05
GPT-6.1 Terra0.7479.72026-11-05
Gemini 4 Carbon0.7379.02026-11-11
Claude Opus 5.50.7177.62026-11-05
Claude Sonnet 5.50.6573.42026-11-05
Gemini 4 Argon0.6472.72026-11-05
Gemini 4 Argon0.6472.72026-10-18
Gemini 4 Argon0.6472.72026-10-17
Claude Opus 50.5969.12026-11-05
GPT-5.6 Sol0.5868.42026-11-05
Claude Fable 5.10.5868.42026-11-05
GPT-5.50.5667.02026-11-05
GPT-6 Sol0.5667.02026-11-05
Claude Fable 50.5264.22026-11-05
Muse Wonder0.5163.52026-10-18
Muse Wonder0.5163.52026-10-17
Muse Wonder0.5163.52026-11-05
DeepSeek V4.1 Flash 10170.4962.12026-11-05
Gemini 3.8 Flash0.4760.72026-11-05
DeepSeek V4.1 Flash0.4357.92026-11-05
DeepSeek V4 Pro 08130.4357.92026-11-05
Qwen3.8 Max0.3854.42026-11-05
GPT-5.40.3753.72026-11-05
Claude Opus 4.80.3652.92026-11-05
Claude Sonnet 50.3552.22026-11-05
GPT-5.6 Terra0.3552.22026-11-05
Gemini 3.1 Pro Preview0.3451.52026-11-05
DeepSeek V4 Flash 07310.3451.52026-11-05
Grok 4.60.3451.52026-11-05
GLM 5.30.3350.82026-11-05
Qwen3.7 Max0.3250.12026-11-05
Gemini 3.6 Flash0.348.72026-11-05
Claude Haiku 5.50.348.72026-11-05
o30.2948.02026-11-05
Grok 4.70.2948.02026-11-05
Claude Opus 4.70.2847.32026-11-05
Gemini 3 Flash Preview0.2645.92026-11-05
Kimi K30.2645.92026-11-05
Claude Opus 4.60.2545.22026-11-05
Muse Spark 1.30.2545.22026-11-05
GPT-50.2343.82026-11-05
GPT-5.20.2343.82026-11-05
Qwen3.6 35B-A3B0.2243.12026-11-05
Claude Opus 4.50.2243.12026-11-05
GPT-5.6 Luna0.2142.42026-11-05
Qwen3.5-Flash0.241.72026-11-05
Mistral Large 40.241.72026-11-05
Qwen3.6 Max Preview0.1941.02026-11-05
Gemini 3.5 Flash Lite0.1941.02026-11-05
GLM-5.20.1941.02026-11-05
Qwen3.5 397B-A17B0.1840.32026-11-05
Qwen3.6 Flash0.1840.32026-11-05
Kimi K2.60.1840.32026-11-05
Qwen3.5 122B-A10B0.1739.62026-11-05
Qwen3.5 Plus0.1739.62026-11-05
Qwen3.7 Plus0.1739.62026-11-05
Claude Sonnet 4.50.1739.62026-11-05
Mistral Large 40.1739.62026-10-17
Mistral Large 40.1739.62026-10-18
Claude Sonnet 4.60.1638.92026-11-05
Qwen3.7 Flash0.1538.22026-11-05
Qwen3.6 Plus0.1236.02026-11-05
GPT-5.4 mini0.1135.32026-11-05
GPT-5 mini0.134.62026-11-05
GPT-5 nano0.0933.92026-11-05
MiniMax-M30.0833.22026-11-05
GLM-5.3 Flash0.0833.22026-10-17
GLM-5.3 Flash0.0833.22026-10-18
GLM-5.3 Flash0.0833.22026-11-05
Qwen3.6 27B0.0732.52026-11-05
o3 Mini High0.0732.52026-11-05
GPT-6 Luna0.0732.52026-11-05
Inkling Small0.0631.82026-11-05
Qwen3 Max0.0531.12026-11-05
o4 Mini High0.0531.12026-11-05
gpt-oss-120b0.0229.02026-11-05

常见问题