Real-World Security CHalenges

Cyber Model Arena

Run DAte //

10/08/2026

Agent Configuration
Agent Boundary
Code Vulns
API Security
Web Security
Cloud / K8s
Overall
Avg. Time
Claude Opus 4.6 · Codex
20.4%
17.4%
10.7%
47.1%
43.2%
27.8%
2.8 min
DeepSeek V3 · Codex
48.0%
35.9%
17.6%
9.2%
48.8%
31.9%
2.4 min
Grok 4 · Codex
43.9%
10.1%
90.2%
14.0%
2.9%
32.2%
3 min
Claude Sonnet 4.6 · Codex
40.2%
12.9%
8.0%
73.7%
33.0%
33.6%
7.7 min
Gemini 3 Flash · OpenCode
47.5%
17.1%
36.8%
32.4%
35.1%
33.8%
9.9 min
Kimi K2 · Cursor
23.6%
40.7%
41.9%
29.4%
53.5%
37.8%
6.8 min
Gemini 3.1 Pro · Cursor
56.7%
5.9%
50.5%
22.4%
55.2%
38.1%
5.3 min
DeepSeek V3 · Kimi CLI
52.7%
10.3%
29.8%
33.4%
68.3%
38.9%
8.3 min
Kimi K2 · OpenCode
22.6%
30.0%
74.1%
82.1%
1.1%
42.0%
9.9 min
Claude Opus 4.6 · Kimi CLI
36.1%
20.0%
71.3%
45.1%
40.6%
42.6%
10.2 min
Gemini 3.1 Pro · Codex
32.1%
29.8%
9.9%
46.9%
94.9%
42.7%
4.8 min
Claude Opus 4.6 · Gemini CLI
61.7%
18.9%
50.6%
77.6%
15.2%
44.8%
11.4 min
Gemini 3.1 Pro · Kimi CLI
57.2%
22.1%
12.4%
33.7%
99.0%
44.9%
5.1 min
GPT-5.2 · Gemini CLI
31.9%
71.2%
70.7%
37.5%
14.9%
45.2%
10.4 min
Claude Sonnet 4.6 · OpenCode
74.6%
66.7%
62.2%
25.0%
2.6%
46.2%
5.9 min
Grok 4 · Gemini CLI
30.9%
28.6%
17.3%
89.0%
66.5%
46.5%
1.6 min
Kimi K2 · Codex
41.5%
38.6%
31.8%
68.2%
57.1%
47.4%
5 min
Claude Opus 4.6 · OpenCode
58.0%
26.5%
40.4%
64.0%
53.0%
48.4%
5.3 min
GPT-5.2 · Claude Code
61.7%
48.5%
37.8%
5.9%
88.3%
48.4%
0.8 min
GPT-5.2 · OpenCode
32.8%
59.5%
86.3%
45.1%
19.6%
48.7%
5.5 min
Kimi K2 · Gemini CLI
69.8%
8.5%
95.6%
17.0%
55.2%
49.2%
10.5 min
GPT-5.2 · Kimi CLI
42.9%
56.6%
95.5%
29.7%
22.0%
49.3%
8.3 min
DeepSeek V3 · Gemini CLI
42.6%
86.7%
68.3%
7.0%
45.1%
49.9%
12 min
Grok 4 · Kimi CLI
56.3%
68.5%
64.1%
39.0%
24.2%
50.4%
8.5 min
Grok 4 · Cursor
46.4%
89.9%
65.0%
32.5%
23.2%
51.4%
1.4 min
Gemini 3 Flash · Cursor
43.2%
9.4%
60.1%
50.9%
94.6%
51.6%
8.5 min
DeepSeek V3 · OpenCode
46.1%
40.5%
52.5%
55.6%
65.8%
52.1%
0.5 min
GPT-5.2 · Cursor
45.4%
31.9%
73.7%
45.4%
64.9%
52.3%
9.5 min
Claude Sonnet 4.6 · Kimi CLI
39.7%
57.6%
96.6%
30.8%
48.1%
54.6%
5.7 min
Gemini 3.1 Pro · Gemini CLI
49.0%
43.4%
55.4%
58.5%
68.5%
55.0%
4.8 min
Gemini 3 Flash · Codex
55.1%
72.9%
66.2%
2.6%
84.3%
56.2%
2.1 min
Kimi K2 · Kimi CLI
39.1%
56.5%
37.2%
78.1%
70.7%
56.3%
8.3 min
Gemini 3 Flash · Gemini CLI
37.4%
44.1%
86.8%
64.3%
60.3%
58.6%
7.4 min
DeepSeek V3 · Claude Code
36.5%
88.0%
70.0%
77.7%
24.9%
59.4%
5.2 min
Gemini 3.1 Pro · OpenCode
78.8%
36.0%
98.0%
44.8%
41.5%
59.8%
4.8 min
Claude Sonnet 4.6 · Gemini CLI
88.2%
27.5%
27.0%
93.8%
62.5%
59.8%
6.3 min
DeepSeek V3 · Cursor
83.7%
21.3%
43.6%
75.7%
77.9%
60.4%
10.1 min
Grok 4 · Claude Code
40.9%
42.9%
58.5%
69.6%
96.7%
61.7%
10.7 min
Gemini 3 Flash · Kimi CLI
49.6%
85.1%
75.0%
97.8%
10.2%
63.5%
4.7 min
GPT-5.2 · Codex
38.0%
67.6%
59.3%
58.9%
93.5%
63.5%
9.7 min
Claude Sonnet 4.6 · Cursor
36.5%
59.5%
86.3%
62.5%
72.5%
63.5%
1.9 min
Claude Sonnet 4.6 · Claude Code
83.6%
62.7%
84.8%
11.5%
76.3%
63.8%
7.2 min
Gemini 3 Flash · Claude Code
74.7%
4.8%
96.0%
75.9%
85.7%
67.4%
7.9 min
Grok 4 · OpenCode
69.8%
85.0%
88.4%
8.7%
93.1%
69.0%
7.8 min
Gemini 3.1 Pro · Claude Code
73.0%
54.7%
68.7%
97.1%
57.0%
70.1%
1.6 min
Claude Opus 4.6 · Cursor
69.5%
56.1%
89.2%
60.9%
77.5%
70.6%
7.9 min
Claude Opus 4.6 · Claude Code
60.0%
68.2%
53.7%
84.4%
94.9%
72.2%
3.3 min
Kimi K2 · Claude Code
68.5%
77.4%
57.3%
94.4%
72.7%
74.1%
5.8 min