← All benchmarks·Source: Epoch AI

ARC-AGI-2 leaderboard

Current leader: GPT-5.6 Sol at 92.5 (% solved).

Cost vs quality

Each dot is a model. The dashed line is the value frontier — top-left is the best trade-off (high score, low cost).

Closed model Open-weight model

Full leaderboard

#ModelScore% solvedOutput $/1MReleased
1
GPT-5.6 Sol
92.5$30.00Jul 2026
2
Claude Opus 5
88.33$50.00Jul 2026
3
GPT-5.5
85$30.00Apr 2026
4
GPT-5.5 Pro
84.58$180.00Apr 2026
5
GPT-5.6 Terra
83.9$7.50Jul 2026
6
GPT-5.4 Pro
83.33$180.00Mar 2026
7
Gemini 3.1 Pro
77.1$12.00Feb 2026
8
Claude Opus 4.7
75.8$150.00Apr 2026
9
GPT-5.4
73.95$15.00Mar 2026
10
Gemini 3.5 Flash
72.08$9.00
11
Claude Opus 4.8
72.08$50.00May 2026
12
Claude Opus 4.6
69.17$25.00Feb 2026
13
Grok 4.20
65.14$2.50Feb 2026
14
Claude Sonnet 4.6
60.42$15.00Feb 2026
15
GPT-5.6 Luna
59.54$3.00Jul 2026
16
GPT-5.2 Pro
54.16Dec 2025
17
GPT-5.2
52.91Dec 2025
18
Claude Opus 4.5
37.64Nov 2025
19
Gemini 3 Flash
33.61Dec 2025
20
Gemini 3 Pro
31.11Nov 2025
21
GLM-5.2OPEN
22.78$2.35Jun 2026
22
GPT-5.4 Mini
18.9$4.50Mar 2026
23
GPT-5 Pro
18.33Oct 2025
24
GPT-5.1
17.64Nov 2025
25
Grok 4
15.97Sep 2025
26
Claude Sonnet 4.5
13.61Sep 2025
27
Kimi K2.5OPEN
11.81$2.85Feb 2026
28
GPT-5
9.86Aug 2025
29
Claude Opus 4
8.61May 2025
30
o3
6.53Dec 2024
31
o4-mini
6.11Apr 2025
32
Claude Sonnet 4
5.93May 2025
33
GPT-5.4 Nano
5.69$1.25Mar 2026
34
Grok 4 Fast
5.28Sep 2025
35
Gemini 2.5 Pro (Jun 2025)
4.86Jun 2025
36
MiniMax-M2.5OPEN
4.86Feb 2026
37
o3-pro
4.86Jun 2025
38
GLM-5OPEN
4.86$2.55Feb 2026
39
GPT-5 mini
4.44Aug 2025
40
Claude Haiku 4.5
4.03Oct 2025
41
DeepSeek-V3.2OPEN
4.03Dec 2025
42
o3-mini
2.99Jan 2025
43
GPT-5 nano
2.61Aug 2025
44
Gemini 2.5 Flash (May 2025)
2.54May 2025
45
DeepSeek-R1OPEN
1.3Jan 2025
46
Gemini 2.0 Flash (Feb 2025)
1.3Dec 2024
47
Qwen3-235B-A22B-Instruct (Jul 2025)OPEN
1.25Jul 2025
48
DeepSeek-R1 (May 2025)OPEN
1.12May 2025
49
Claude 3.7 Sonnet
0.9Feb 2025
50
o1-mini
0.83Sep 2024
51
GPT-4.5
0.8Feb 2025
52
Gemini 1.5 Pro (Sept 2024)
0.8Sep 2024
53
Grok-3 mini
0.42Feb 2025
54
GPT-4.1
0.42Apr 2025
55
Grok 3
0Feb 2025
56
Llama 4 MaverickOPEN
0Apr 2025
57
Llama 4 ScoutOPEN
0Apr 2025
58
Magistral Small 1.1OPEN
0Jun 2025
59
GPT-4.1 mini
0Apr 2025
60
GPT-4.1 nano
0Apr 2025
61
GPT-4o (Nov 2024)
0Nov 2024
62
GPT-4o mini
0Jul 2024

Scores from Epoch AI, updated . Cost is joined from our daily API pricing where a model matches; models we don't price still appear with their score. Always verify on the benchmark's own site before relying on a number.