Benchgen

CyberGym — Results

RankModelScore
1deepseek-v4-1-flash88.1
2deepseek-v4-1-flash88.1
3gpt-5-5-cyber85.6
4gpt-5-6-sol84.5
5glm-5-384.5
6deepseek-v4-pro83.3
7claude-mythos-preview83.1
8gpt-5-581.8
9gpt-5-479
10hy4-preview78.4
11deepseek-v4-flash-vision-exp75.3
12claude-opus-4-773.1
13seed-2-1-pro68.7
14glm-5-168.7
15seed-2-1-turbo67
16claude-opus-4-666.6
17claude-sonnet-4-665.2
18claude-opus-4-550.6
19muse-spark43.5
20glm-543.2
21kimi-k2-541.3
22gpt-539.4
23gemini-3-1-pro38.8
24claude-sonnet-4-528.9
25claude-opus-4-125
C

CyberGym — Evaluations

1 phaseActive

No evaluations yet for this benchmark.