ARC-AGI-2 · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	GPT-6 Astra	max	95.0%	$1.12	–
2	GPT-6.1 Sol	max	94.2%	$0.254	–
3	Opus 5.5	high	93.3%	$0.408	–
4	GPT-5.6 Sol	max	92.5%	$1.44	–
5	Gemini 4	high	91.3%	$0.325	–
6	Opus 5	max	90.4%	$2.06	–
7	Fable 5.1	max	90.0%	$4.49	–
8	GPT-6 Sol	max	89.6%	$0.439	–
9	Fable 5	max	89.2%	$5.45	–
10	Gemini 3.8 Flash	high	89.2%	$0.400	–
11	GPT-5.5	xhigh	85.0%	$1.87	–
12	Gemini 3.7 Flash	high	84.6%	$0.249	–
13	Gemini 3	–	84.6%	$13.62	–
14	GPT-5.5 Pro	high	84.6%	$10.51	–
15	GPT-5.6 Terra	max	83.9%	$1.09	–
16	GPT-5.4 Pro	xhigh	83.3%	$16.41	–
17	Gemini 3.1 Pro	high	77.1%	$0.962	–
18	Opus 4.7	max	75.8%	$7.43	–
19	GPT-5.4	xhigh	74.0%	$1.52	–
20	DeepSeek V4.1 Flash	max	72.9%	$0.129	–
21	Opus 4.8	high	72.1%	$2.74	–
22	Gemini 3.5 Flash	high	72.1%	$0.850	–
23	Opus 4.6	high	69.2%	$3.47	–
24	Grok 4.6	xhigh	67.1%	$0.757	–
25	GLM 5.3 Flash	max	65.8%	$0.093	–
26	Grok 4.20	xhigh	65.1%	$0.920	–
27	DeepSeek V4 Flash 0731	max	61.4%	$0.042	–
28	Grok 4.7	xhigh	61.4%	$2.01	–
29	DeepSeek V4 Pro 0813	max	61.3%	$0.598	–
30	Haiku 5.5	xhigh	61.3%	$0.085	–
31	Sonnet 4.6	high	60.4%	$2.70	–
32	Kimi K3	max	60.4%	$1.59	–
33	Gemini 3.6 Flash	high	60.4%	$0.606	–
34	GPT-5.6 Luna	max	59.6%	$0.177	–
35	GPT-6 Luna	max	59.3%	$0.062	–
36	GPT-5.2 Pro	high	54.2%	$15.72	–
37	GPT-5.2	xhigh	52.9%	$1.90	–
38	Grok 4.5	high	52.6%	$0.780	–
39	Qwen 3.8 27B	xhigh	42.4%	$0.447	–
40	Inkling Small	xhigh	40.1%	$0.233	–
41	Opus 4.5	high	37.6%	$2.40	–
42	Inkling	max	36.5%	$0.642	–
43	Gemini 3 Flash	high	33.6%	$0.231	–
44	Gemini 3 Pro	high	31.1%	$0.811	–
45	GLM 5.2	max	22.8%	$0.250	–
46	GPT-5.4 Mini	xhigh	18.9%	$0.750	–
47	GPT-5 Pro	high	18.3%	$7.14	–
48	GPT-5.1	high	17.6%	$1.17	–
49	Grok 4	high	16.0%	$2.17	–
50	Sonnet 4.5	high	13.6%	$0.759	–
51	Kimi K2.5	max	11.8%	$0.280	–
52	Gemini 3.5 Flash-Lite	high	10.3%	$0.139	–
53	GPT-5	high	9.9%	$0.730	–
54	Opus 4	medium	8.6%	$1.93	–
55	o3	high	6.5%	$0.834	–
56	o4 mini	high	6.1%	$0.856	–
57	Sonnet 4	high	5.9%	$0.486	–
58	GPT-5.4 Nano	xhigh	5.7%	$0.160	–
59	Grok 4 Fast	none	5.3%	$0.061	–
60	Gemini 2.5 Pro	high	4.9%	$0.757	–
61	MiniMax M2.5	high	4.9%	$0.170	–
62	GLM 5	high	4.9%	$0.270	–
63	GPT-5 Mini	high	4.4%	$0.198	–
64	Haiku 4.5	xhigh	4.0%	$0.377	–
65	DeepSeek V3.2	high	4.0%	$0.120	–
66	o3 mini	high	3.0%	$0.547	–
67	GPT-5 Nano	high	2.6%	$0.029	–
68	Gemini 2.5 Flash	medium	2.5%	$0.319	–
69	Gemini 2.0 Flash	medium	1.3%	$0.0040	–
70	DeepSeek R1	xhigh	1.3%	$0.080	–
71	Qwen 3 235B-A22B	xhigh	1.3%	$0.0044	–
72	o1 mini	high	0.8%	$0.191	–
73	Gemini 1.5 Pro	none	0.8%	$0.040	–
74	GPT-4.5	none	0.8%	$2.10	–
75	GPT-4.1	high	0.4%	$0.069	–
76	Grok 3 Mini	low	0.4%	$0.013	–
77	GPT-4o	high	0.0%	$0.080	–
78	GPT-4o mini	high	0.0%	$0.010	–
79	Llama 4 Maverick	medium	0.0%	$0.012	–
80	Llama 4 Scout	medium	0.0%	$0.0062	–
81	GPT-4.1 Nano	high	0.0%	$0.0036	–
82	GPT-4.1 Mini	high	0.0%	$0.014	–
83	Grok 3	none	0.0%	$0.142	–
84	Magistral Small	–	0.0%	$0.049	–
85	Magistral Medium	–	0.0%	$0.108	–