ARC-AGI-2

Abstract visual puzzles that are easy for people and hard for AI. Published by the ARC Prize Foundation.

CategoryReasoning
MetricAccuracy
SourceEpoch AI (evaluator)
LicenseCC BY 4.0
Benchmark pagearcprize.org
StatusScored
Weight1.3 within Reasoning
FitDiscrimination 2.57, noise 0.49

Results

#ModelProviderResultDate
1 GPT-6 Astra Max OpenAI 95.0% 2026-11-11
2 GPT-6.1 Sol Max OpenAI 94.2% 2026-11-11
3 Gemini 4 Carbon Medium Google 94.2% 2026-11-11
4 Claude Fable 5.2 Extra High Anthropic 93.8% 2026-11-11
5 GPT-6 Astra Extra High OpenAI 93.3% 2026-11-11
6 Claude Opus 5.5 High Anthropic 93.3% 2026-11-11
7 Claude Fable 5.2 Max Anthropic 92.5% 2026-11-11
8 GPT-5.6 Sol Max OpenAI 92.5% 2026-11-11
9 Claude Opus 5.5 Extra High Anthropic 92.5% 2026-11-11
10 Gemini 4 Carbon High Google 92.1% 2026-11-11
11 GPT-6 Astra Medium OpenAI 92.1% 2026-11-11
12 GPT-6 Astra High OpenAI 92.1% 2026-11-11
13 Claude Opus 5.5 Max Anthropic 91.7% 2026-11-11
14 GPT-6.1 Sol High OpenAI 91.7% 2026-11-11
15 GPT-6.1 Sol Extra High OpenAI 91.7% 2026-11-11
16 Gemini 4 Argon High Google 91.3% 2026-11-11
17 Claude Opus 5 Max Anthropic 90.4% 2026-11-11
18 Claude Fable 5.2 High Anthropic 90.4% 2026-11-11
19 Claude Fable 5.1 Max Anthropic 90.0% 2026-11-11
20 Claude Fable 5.1 Extra High Anthropic 90.0% 2026-11-11
21 GPT-5.6 Sol Extra High OpenAI 90.0% 2026-11-11
22 GPT-6 Sol Max OpenAI 89.6% 2026-11-11
23 Claude Fable 5.2 Medium Anthropic 89.6% 2026-11-11
24 Gemini 3.8 Flash High Google 89.2% 2026-11-11
25 Claude Fable 5 Max Anthropic 89.2% 2026-11-11
26 Claude Fable 5.1 High Anthropic 88.8% 2026-11-11
27 Claude Fable 5 Extra High Anthropic 88.3% 2026-11-11
28 Claude Opus 5 High Anthropic 88.3% 2026-11-11
29 Claude Fable 5 High Anthropic 87.5% 2026-11-11
30 Claude Opus 5.5 Medium Anthropic 87.5% 2026-11-11
31 Gemini 4 Argon Medium Google 87.1% 2026-11-11
32 GPT-6.1 Sol Medium OpenAI 86.7% 2026-11-11
33 Claude Fable 5.2 Low Anthropic 86.7% 2026-11-11
34 Claude Fable 5.1 Medium Anthropic 86.3% 2026-11-11
35 GPT-6 Astra Low OpenAI 85.4% 2026-11-11
36 GPT-5.6 Sol High OpenAI 85.4% 2026-11-11
37 GPT-5.5 Extra High OpenAI 85.0% 2026-11-11
38 Gemini 3 Deep Think Google 84.6% 2026-11-11
39 Gemini 3.7 Flash High Google 84.6% 2026-11-11
40 GPT-5.5 Pro High OpenAI 84.6% 2026-11-11
41 GPT-5.5 Pro Extra High OpenAI 84.2% 2026-11-11
42 GPT-5.6 Terra Max OpenAI 83.9% 2026-11-11
43 Claude Sonnet 5.5 Extra High Anthropic 83.9% 2026-11-11
44 GPT-5.4 Pro Extra High OpenAI 83.3% 2026-11-11
45 GPT-5.5 High OpenAI 83.3% 2026-11-11
46 Gemini 3.8 Flash Medium Google 82.9% 2026-11-11
47 Claude Fable 5 Medium Anthropic 82.5% 2026-11-11
48 Inkling 2 Small Extra High Thinking Machines 81.7% 2026-11-11
49 Claude Fable 5.1 Low Anthropic 78.3% 2026-11-11
50 DeepSeek V4.1 Flash 1017 Max DeepSeek 78.1% 2026-11-11
51 GPT-6 Sol Extra High OpenAI 78.1% 2026-11-11
52 Gemini 3.8 Flash Low Google 77.5% 2026-11-11
53 Gemini 3.1 Pro Preview Google 77.1% 2026-11-11
54 Claude Fable 5 Low Anthropic 76.8% 2026-11-11
55 GPT-6.1 Sol Low OpenAI 76.7% 2026-11-11
56 Claude Opus 4.7 Max Anthropic 75.8% 2026-11-11
57 GPT-5.6 Terra Extra High OpenAI 74.2% 2026-11-11
58 GPT-5.4 Extra High OpenAI 74.0% 2026-11-11
59 DeepSeek V4.1 Flash 1017 Low DeepSeek 73.6% 2026-11-11
60 DeepSeek V4.1 Flash Max DeepSeek 72.9% 2026-11-11
61 Gemini 3.5 Flash High Google 72.1% 2026-11-11
62 Claude Opus 4.8 High Anthropic 72.1% 2026-11-11
63 Claude Opus 4.8 Medium Anthropic 71.7% 2026-11-11
64 GPT-5.5 Medium OpenAI 70.4% 2026-11-11
65 Claude Opus 5.5 Low Anthropic 70.1% 2026-11-11
66 Claude Opus 4.6 Thinking Anthropic 69.2% 2026-11-11
67 GPT-6 Sol High OpenAI 68.9% 2026-11-11
68 DeepSeek V4.1 Flash 1017 High DeepSeek 68.6% 2026-11-11
69 Claude Opus 4.7 High Anthropic 68.3% 2026-11-11
70 GPT-5.4 High OpenAI 67.5% 2026-11-11
71 DeepSeek V4.1 Flash High DeepSeek 67.5% 2026-11-11
72 Claude Opus 4.7 Medium Anthropic 67.5% 2026-11-11
73 Claude Sonnet 5.5 High Anthropic 67.4% 2026-11-11
74 Grok 4.6 Extra High xAI 67.1% 2026-11-11
75 GPT-5.6 Sol Medium OpenAI 67.1% 2026-11-11
76 GPT-5.6 Terra High OpenAI 67.1% 2026-11-11
77 GLM-5.3-Flash Max Zhipu AI 65.8% 2026-11-11
78 Grok 4.20 (Reasoning) xAI 65.1% 2026-11-11
79 Grok 4.6 High xAI 65.1% 2026-11-11
80 Gemini 3.7 Flash Medium Google 63.8% 2026-11-11
81 Claude Opus 4.8 Low Anthropic 62.2% 2026-11-11
82 Claude Opus 4.7 Low Anthropic 62.1% 2026-11-11
83 Grok 4.7 Extra High xAI 61.4% 2026-11-11
84 DeepSeek V4 Flash 0731 Max DeepSeek 61.4% 2026-11-11
85 DeepSeek V4 Pro 0813 Max DeepSeek 61.3% 2026-11-11
86 Claude Haiku 5.5 Extra High Anthropic 61.3% 2026-11-11
87 Grok 4.6 Medium xAI 61.3% 2026-11-11
88 DeepSeek V4.1 Flash Low DeepSeek 60.6% 2026-11-11
89 Gemini 3.6 Flash High Google 60.4% 2026-11-11
90 Kimi K3 Max Moonshot AI 60.4% 2026-11-11
91 Claude Sonnet 4.6 High Anthropic 60.4% 2026-11-11
92 DeepSeek V4 Pro 0813 High DeepSeek 59.7% 2026-11-11
93 GPT-6 Astra Non-reasoning OpenAI 59.6% 2026-11-11
94 GPT-5.6 Luna Max OpenAI 59.5% 2026-11-11
95 GPT-6 Luna Max OpenAI 59.3% 2026-11-11
96 Grok 4.7 Medium xAI 58.8% 2026-11-11
97 Claude Sonnet 4.6 Max Anthropic 58.3% 2026-11-11
98 Grok 4.7 High xAI 58.3% 2026-11-11
99 GPT-6 Sol Medium OpenAI 57.8% 2026-11-11
100 Muse Spark Open Extra High Meta 57.5% 2026-11-11
101 DeepSeek V4 Pro 0813 Low DeepSeek 56.3% 2026-11-11
102 DeepSeek V4 Flash 0731 High DeepSeek 56.0% 2026-11-11
103 GPT-5.4 Medium OpenAI 55.4% 2026-11-11
104 Kimi K3 High Moonshot AI 55.0% 2026-11-11
105 GPT-5.2 Pro High OpenAI 54.2% 2026-11-11
106 Gemini 3.7 Flash Low Google 52.9% 2026-11-11
107 GPT-5.2 Extra High OpenAI 52.9% 2026-11-11
108 Grok 4.5 High xAI 52.6% 2026-11-11
109 Grok 4.5 Medium xAI 52.6% 2026-11-11
110 Gemini 3.6 Flash Medium Google 50.4% 2026-11-11
111 GLM-5.3-Flash High Zhipu AI 50.1% 2026-11-11
112 Gemini 4 Carbon Low Google 49.6% 2026-11-11
113 GPT-5.6 Luna Extra High OpenAI 47.6% 2026-11-11
114 Muse Spark Open High Meta 47.5% 2026-11-11
115 DeepSeek V4 Flash 0731 Low DeepSeek 46.0% 2026-11-11
116 GPT-5.2 High OpenAI 43.3% 2026-11-11
117 GPT-5.6 Sol Low OpenAI 42.5% 2026-11-11
118 Qwen3.8 27B Extra High Alibaba 42.4% 2026-11-11
119 Gemini 4 Argon Low Google 42.1% 2026-11-11
120 GPT-6 Luna Extra High OpenAI 41.9% 2026-11-11
121 Inkling Small Extra High Thinking Machines 40.1% 2026-11-11
122 GPT-5.2 Pro Medium OpenAI 38.5% 2026-11-11
123 Claude Opus 4.5 Thinking Anthropic 37.6% 2026-11-11
124 GPT-5.6 Terra Medium OpenAI 37.5% 2026-11-11
125 Inkling Thinking Machines 36.5% 2026-11-11
126 Gemini 3 Flash Preview Google 33.6% 2026-11-11
127 GPT-5.5 Low OpenAI 33.3% 2026-11-11
128 Grok 4.5 Low xAI 33.1% 2026-11-11
129 Inkling Small High Thinking Machines 33.1% 2026-11-11
130 Claude Sonnet 5.5 Medium Anthropic 32.8% 2026-11-11
131 GPT-6 Sol Low OpenAI 31.5% 2026-11-11
132 GPT-6 Luna High OpenAI 31.4% 2026-11-11
133 Gemini 3 Pro Preview Google 31.1% 2026-11-11
134 Claude Haiku 5.5 High Anthropic 30.7% 2026-11-11
135 Gemini 3.6 Flash Low Google 30.4% 2026-11-11
136 GPT-5.6 Luna High OpenAI 29.3% 2026-11-11
137 GPT-5.4 Low OpenAI 29.2% 2026-11-11
138 GLM-5.3-Flash Low Zhipu AI 27.9% 2026-11-11
139 Grok 4.6 Low xAI 27.6% 2026-11-11
140 Muse Spark Open Medium Meta 27.5% 2026-11-11
141 GPT-5.2 Medium OpenAI 26.7% 2026-11-11
142 Claude Sonnet 5.5 Low Anthropic 24.0% 2026-11-11
143 Beam 1.0 Extra High Reflection 23.5% 2026-11-11
144 GLM-5.2 Zhipu AI 22.8% 2026-11-11
145 Qwen3.8 27B Low Alibaba 22.8% 2026-11-11
146 Beam 1.0 High Reflection 20.0% 2026-11-11
147 GPT-5.4 mini Extra High OpenAI 18.9% 2026-11-11
148 GPT-5.6 Terra Low OpenAI 18.8% 2026-11-11
149 GPT-5 Pro OpenAI 18.3% 2026-11-11
150 GPT-6 Luna Medium OpenAI 18.1% 2026-11-11
151 GPT-5.1 High OpenAI 17.6% 2026-11-11
152 Grok 4.7 Low xAI 16.7% 2026-11-11
153 Grok 4 xAI 16.0% 2026-11-11
154 Claude Sonnet 4.5 Thinking Anthropic 13.6% 2026-11-11
155 Inkling Small Medium Thinking Machines 13.6% 2026-11-11
156 GPT-5.4 mini High OpenAI 13.2% 2026-11-11
157 Qwen3.8 27B Medium Alibaba 13.2% 2026-11-11
158 Kimi K3 Low Moonshot AI 12.4% 2026-11-11
159 Muse Spark Open Low Meta 12.2% 2026-11-11
160 Kimi K2.5 Moonshot AI 11.8% 2026-11-11
161 Gemini 3.5 Flash Lite High Google 10.3% 2026-11-11
162 GPT-5 High OpenAI 9.9% 2026-11-11
163 GPT-5.2 Low OpenAI 9.7% 2026-11-11
164 Claude Haiku 5.5 Medium Anthropic 9.4% 2026-11-11
165 Gemini 3.5 Flash Minimal Google 8.9% 2026-11-11
166 Claude Opus 4 Thinking Anthropic 8.6% 2026-11-11
167 Claude Opus 4.5 Anthropic 7.8% 2026-11-11
168 GPT-5 Medium OpenAI 7.5% 2026-11-11
169 GPT-5.6 Luna Medium OpenAI 7.4% 2026-11-11
170 GPT-5.1 Medium OpenAI 6.5% 2026-11-11
171 o3 High OpenAI 6.5% 2026-11-11
172 Beam 1.0 Medium Reflection 6.4% 2026-11-11
173 tiny-recursion-model Unknown 6.3% 2026-11-11
174 o4-mini High OpenAI 6.1% 2026-11-11
175 Claude Sonnet 4 Thinking Anthropic 5.9% 2026-11-11
176 GPT-5.4 nano Extra High OpenAI 5.7% 2026-11-11
177 Grok 4 Fast xAI 5.3% 2026-11-11
178 Gemini 3.5 Flash Lite Medium Google 5.3% 2026-11-11
179 GPT-5.6 Luna Low OpenAI 5.1% 2026-11-11
180 MiniMax-M2.5 MiniMax 4.9% 2026-11-11
181 GLM-5 Zhipu AI 4.9% 2026-11-11
182 o3-pro High OpenAI 4.9% 2026-11-11
183 Inkling Small Low Thinking Machines 4.9% 2026-11-11
184 Gemini 2.5 Pro Thinking Google 4.9% 2026-11-11
185 GPT-6 Luna Low OpenAI 4.6% 2026-11-11
186 GPT-5 Mini High OpenAI 4.4% 2026-11-11
187 GPT-5.4 mini Medium OpenAI 4.4% 2026-11-11
188 DeepSeek V3.2 DeepSeek 4.0% 2026-11-11
189 Claude Haiku 4.5 Thinking Anthropic 4.0% 2026-11-11
190 GPT-5 Mini Medium OpenAI 4.0% 2026-11-11
191 Inkling Small Minimal Thinking Machines 4.0% 2026-11-11
192 Claude Sonnet 4.5 Anthropic 3.8% 2026-11-11
193 GPT-5.4 nano High OpenAI 3.6% 2026-11-11
194 o3-mini High OpenAI 3.0% 2026-11-11
195 o3 Medium OpenAI 3.0% 2026-11-11
196 Claude Haiku 5.5 Low Anthropic 2.8% 2026-11-11
197 Beam 1.0 Low Reflection 2.8% 2026-11-11
198 Gemini 3.6 Flash Minimal Google 2.6% 2026-11-11
199 GPT-5 Nano High OpenAI 2.6% 2026-11-11
200 Gemini 2.5 Flash Preview 05-20 Thinking Google 2.5% 2026-11-11
201 o4-mini Medium OpenAI 2.4% 2026-11-11
202 o3-mini Medium OpenAI 2.1% 2026-11-11
203 DeepSeek V4.1 Flash 1017 Non-reasoning DeepSeek 2.1% 2026-11-11
204 DeepSeek V4 Flash 0731 Non-reasoning DeepSeek 2.1% 2026-11-11
205 o3-pro Low OpenAI 2.1% 2026-11-11
206 o3 Low OpenAI 2.0% 2026-11-11
207 GPT-5.1 Low OpenAI 1.9% 2026-11-11
208 GPT-5 Low OpenAI 1.9% 2026-11-11
209 GPT-5.4 nano Medium OpenAI 1.9% 2026-11-11
210 o3-pro Medium OpenAI 1.9% 2026-11-11
211 Gemini 2.5 Flash Preview 05-20 Google 1.7% 2026-11-11
212 GPT-5 Mini Minimal OpenAI 1.7% 2026-11-11
213 o4-mini Low OpenAI 1.7% 2026-11-11
214 GPT-6 Sol Non-reasoning OpenAI 1.7% 2026-11-11
215 GPT-5.4 nano Low OpenAI 1.5% 2026-11-11
216 Gemini 3.5 Flash Lite Low Google 1.5% 2026-11-11
217 Qwen3.8 27B Non-reasoning Alibaba 1.5% 2026-11-11
218 DeepSeek-R1 DeepSeek 1.3% 2026-11-11
219 Gemini 2.0 Flash Google 1.3% 2026-11-11
220 Claude Opus 4 Anthropic 1.3% 2026-11-11
221 Claude Sonnet 4 Anthropic 1.3% 2026-11-11
222 codex-mini-2025-05-16 Unknown 1.3% 2026-11-11
223 Claude Haiku 4.5 Anthropic 1.3% 2026-11-11
224 Qwen3 235B-A22B Instruct 2507 Alibaba 1.3% 2026-11-11
225 DeepSeek V4.1 Flash Non-reasoning DeepSeek 1.3% 2026-11-11
226 DeepSeek R1 0528 DeepSeek 1.1% 2026-11-11
227 GPT-5.4 mini Low OpenAI 1.1% 2026-11-11
228 Claude Sonnet 3.7 Thinking Anthropic 0.9% 2026-11-11
229 GPT-5 Nano Medium OpenAI 0.9% 2026-11-11
230 GPT-5.2 OpenAI 0.8% 2026-11-11
231 o1-mini OpenAI 0.8% 2026-11-11
232 DeepSeek V4 Pro 0813 Non-reasoning DeepSeek 0.8% 2026-11-11
233 Gemini 3.5 Flash Lite Minimal Google 0.8% 2026-11-11
234 GPT-5 Mini Low OpenAI 0.8% 2026-11-11
235 Gemini 1.5 Pro (Sept 2024) Google 0.8% 2026-11-11
236 GPT-4.5 OpenAI 0.8% 2026-11-11
237 GPT-4.1 OpenAI 0.4% 2026-11-11
238 GPT-5.1 Non-reasoning OpenAI 0.4% 2026-11-11
239 Grok-3 mini Low xAI 0.4% 2026-11-11
240 Magistral Small Mistral 0.0% 2026-11-11
241 Claude Sonnet 3.7 Anthropic 0.0% 2026-11-11
242 GPT-4.1 mini OpenAI 0.0% 2026-11-11
243 GPT-4o (2024-11-20) OpenAI 0.0% 2026-11-11
244 GPT-4o mini OpenAI 0.0% 2026-11-11
245 GPT-4.1 nano OpenAI 0.0% 2026-11-11
246 Llama 4 Maverick Meta 0.0% 2026-11-11
247 Llama 4 Scout Meta 0.0% 2026-11-11
248 Grok 3 xAI 0.0% 2026-11-11
249 Magistral Medium 1.1 Mistral 0.0% 2026-11-11
250 GPT-5 Minimal OpenAI 0.0% 2026-11-11
251 o3-mini Low OpenAI 0.0% 2026-11-11
252 GPT-5 Nano Low OpenAI 0.0% 2026-11-11
253 GPT-5 Nano Minimal OpenAI 0.0% 2026-11-11
254 GPT-6 Luna Non-reasoning OpenAI 0.0% 2026-11-11
255 Gemini 2.5 Pro Preview 06-05 Thinking Google 0.0% 2026-11-11
256 Inkling Small Non-reasoning Thinking Machines 0.0% 2026-11-11