Pass IndexThe State of AISign in

ARC-AGI-2 (Public Eval)

ARC-AGI-2 (Public Eval) is run by ARC Prize Foundation. It has ranked 282 models, of which the catalogue holds 75, scoring from 0 to 99.17.

measured by ARC Prize Foundation · the board itself

PlaceModelMetricScore
2ndof 282Claude Fable 5.1Score (%) (xhigh)99.17
4thof 282GPT 6 AstraScore (%) (xhigh)97.92
6thof 282Claude Opus 5.5Score (%) (max)97.92
9thof 282GPT 6.1 SolScore (%) (max)97.08
10thof 282Gemini 4 ArgonScore (%) (high)97.08
11thof 282Claude Opus 5Score (%) (max)97.06
13thof 282Claude Fable 5Score (%) (max)96.39
19thof 282GPT-5.6 SolScore (%) (max)93.75
24thof 282Gemini 3.8 FlashScore (%) (high)92.5
25thof 282GPT-5.4 ProScore (%) (xhigh)92.21
28thof 282GPT-5.6 TerraScore (%) (max)91.32
31stof 282GPT-5.5 ProScore (%) (xhigh)90.42
33rdof 282Claude Sonnet 5.5Score (%) (xhigh)90.14
34thof 282GPT 6 SolScore (%) (max)90
35thof 282Gemini 3.1 Pro PreviewScore (%)88.07
41stof 282Gemini 3.7 FlashScore (%) (high)85.28
42ndof 282GPT-5.4Score (%) (xhigh)84.17
43rdof 282DeepSeek V4.1 Flash 1017Score (%) (high)83.06
45thof 282Claude Opus 4.7Score (%) (max)82.23
49thof 282DeepSeek V4.1 FlashScore (%) (max)81.67
53rdof 282Claude Opus 4.6Score (%) (high)79.03
61stof 282Gemini 3.5 FlashScore (%) (high)72.08
68thof 282Grok 4.6Score (%) (xhigh)68.33
72ndof 282GLM 5.3 FlashScore (%) (max)66.81
74thof 282Claude Sonnet 4.6Score (%) (high)65.69
75thof 282Gemini 3.6 FlashScore (%) (high)65.28
77thof 282Claude Haiku 5.5Score (%) (xhigh)64.44
78thof 282GPT-5.6 LunaScore (%) (max)64.17
79thof 282DeepSeek V4 Flash (0731)Score (%) (max)63.89
81stof 282DeepSeek V4 Pro 0813Score (%) (high)63.61
84thof 282GPT 6 LunaScore (%) (max)61.81
85thof 282Kimi K3Score (%) (max)61.67
87thof 282Grok 4.7Score (%) (xhigh)60
94thof 282Grok 4.5Score (%) (high)58.19
98thof 282Muse Spark OpenScore (%) (xhigh)57.22
104thof 282GPT-5.2 ProScore (%) (high)51.68
112thof 282Inkling SmallScore (%) (xhigh)39.31
115thof 282Qwen3.8 27BScore (%) (xhigh)38.75
118thof 282InklingScore (%)37.78
139thof 282GLM 5.2Score (%)20.83
143rdof 282GPT 5.1 ThinkingScore (%) (high)18.33
144thof 282GPT-5.4 miniScore (%) (xhigh)17.78
153rdof 282GPT-5 ProScore (%)13.33
154thof 282Kimi K2.5Score (%)12.08
158thof 282GPT-5Score (%) (high)9.58
162ndof 282Gemini 3.5 Flash-LiteScore (%) (high)7.92
166thof 282o4-miniScore (%) (high)7.52
172ndof 282GPT-5 miniScore (%) (high)5.83
173rdof 282MiniMax M2.5Score (%)5.42
174thof 282GLM-5Score (%)5.42
178thof 282GPT-5.4 nanoScore (%) (high)5.14
185thof 282o3Score (%) (medium)4.49
186thof 282Claude Sonnet 4Score (%)4.44
190thof 282o3-proScore (%) (high)3.89
191stof 282DeepSeek v3.2Score (%)3.89
192ndof 282Claude Sonnet 4.5Score (%)3.75
197thof 282Claude Opus 4Score (%)3.33
216thof 282Gemini 2.5 FlashScore (%)1.94
219thof 282Nemotron 3.5 Super VL 120B A12BScore (%) (high)1.94
227thof 282Gemini 2.5 ProScore (%)1.11
229thof 282Qwen3-235B-A22B-Instruct-2507Score (%)83
238thof 282GPT-5 nanoScore (%) (high)30
239thof 282DeepSeek-R1 (0528)Score (%)29
246thof 282GPT-4o (2024-11-20)Score (%)0
247thof 282GPT-4o-mini (2024-07-18)Score (%)0
250thof 282GPT-4.1 nanoScore (%)0
251stof 282GPT-4.1 miniScore (%)0
252ndof 282GPT-4.1 fine-tunedScore (%)0
253rdof 282o1-miniScore (%)0
254thof 282o3-miniScore (%) (low)0
258thof 282Codex miniScore (%)0
259thof 282Grok 3Score (%)0
260thof 282Grok 3 MiniScore (%) (low)0
261stof 282Magistral Small 2506Score (%)0
269thof 282Claude Haiku 4.5Score (%)0

Read from the board on 2026-10-31