Skip to content

Benchmark

ARC-AGI

Scores are percentages.

Models measured
102
Scores recorded
295
Sources
2
IQ weight
2.26

Score distribution

64.1%
Mean
75%
Median
31.2%
Std dev
14.5%
10th percentile
97%
90th percentile
0%
Min
0%98.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.26 for chat models, measured over 102 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
101Muse Spark Open 816B A42Bhigh· Epoch AI Benchmarking Hub data88.5%
102Claude Opus 4.8low· Epoch AI Benchmarking Hub data88%
103GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data88%
104OpenAI o3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim88%
105GPT-5.6 Lunaxhigh· Epoch AI Benchmarking Hub data87.7%
106Claude Haiku 5.5xhigh· Epoch AI Benchmarking Hub data87.5%
107Grok 4.6medium· Epoch AI Benchmarking Hub data87.5%
108Qwen 3.8 27Bxhigh· Epoch AI Benchmarking Hub data87.5%
109Grok 4.5medium· Epoch AI Benchmarking Hub data87.2%
110DeepSeek V4 Pro 0813high· Epoch AI Benchmarking Hub data87.2%
111DeepSeek V4 Flash 0731high· Epoch AI Benchmarking Hub data87%
112Grok 4.6high· Epoch AI Benchmarking Hub data87%
113Grok 4.6xhigh· Epoch AI Benchmarking Hub data87%
114GPT-6 Lunamax· Epoch AI Benchmarking Hub data86.7%
115Kimi K3high· Epoch AI Benchmarking Hub data86.7%
116Claude Sonnet 4.6high· Epoch AI Benchmarking Hub data86.5%
117GPT 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
86.2%
118GPT 5.2xhigh· Epoch AI Benchmarking Hub data86.2%
119GPT 5.4medium· Epoch AI Benchmarking Hub data86.2%
120Claude Sonnet 4.6max· Epoch AI Benchmarking Hub data86%
121GPT-6 Astranone· Epoch AI Benchmarking Hub data86%
122GPT-5.2 Prohigh· Epoch AI Benchmarking Hub data85.7%
123Grok 4.5high· Epoch AI Benchmarking Hub data85.7%
124Gemini 4 Carbonlow· Epoch AI Benchmarking Hub data85.5%
125Gemini 3.7 Flashlow· Epoch AI Benchmarking Hub data85.2%
126Gemini 3 Flash· Epoch AI Benchmarking Hub data84.7%
127Inkling 2 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.5%
128DeepSeek V4 Flash 0731low· Epoch AI Benchmarking Hub data84%
129GPT-6.2 Sollow· Epoch AI Benchmarking Hub data84%
130Inkling Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84%
131Inkling Smallxhigh· Epoch AI Benchmarking Hub data84%
132GPT-6 Solmedium· Epoch AI Benchmarking Hub data83.7%
133Gemini 3.6 Flashmedium· Epoch AI Benchmarking Hub data83.2%
134Claude Haiku 5.5high· Epoch AI Benchmarking Hub data81.5%
135GPT-5.2 Promedium· Epoch AI Benchmarking Hub data81.2%
136Claude Opus 4.564k· Epoch AI Benchmarking Hub data80%
137Inkling FP4· Epoch AI Benchmarking Hub data79.5%
138Grok 4.5low· Epoch AI Benchmarking Hub data79.2%
139GPT 5.2high· Epoch AI Benchmarking Hub data78.7%
140Inkling Smallhigh· Epoch AI Benchmarking Hub data78%
141Muse Spark Open 816B A42Bmedium· Epoch AI Benchmarking Hub data77.7%
142GLM 5.2· Epoch AI Benchmarking Hub data77%
143GPT-5.6 Terramedium· Epoch AI Benchmarking Hub data77%
144Gemini 3.6 Flashlow· Epoch AI Benchmarking Hub data76.5%
145Gemini 4 Argonlow· Epoch AI Benchmarking Hub data76.5%
146GPT-5.6 Lunahigh· Epoch AI Benchmarking Hub data76.5%
147GPT 5.5low· Epoch AI Benchmarking Hub data76.2%
148Claude Opus 4.532k· Epoch AI Benchmarking Hub data75.8%
149Gemini 3 Pro Preview· Epoch AI Benchmarking Hub data75%
150Grok 4.6low· Epoch AI Benchmarking Hub data74.8%

Showing 101–150 of 295 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · Data by LLM Stats