Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Software Engineering on SWE-Bench Verified (ACC & ASR)
Loading...
70.6
Accuracy
Non-attack
55.312
59.281
63.25
67.219
Apr 8, 2026
Accuracy
Average Success Rate (ASR)
Updated 9d ago
Evaluation Results
Method
Method
Links
Accuracy
Average Success Rate (ASR)
Non-attack
Backbone Model=GPT-5.2...
2026.04
70.6
0
SkillTrojan
Backbone Model=GPT-5.2...
2026.04
70.4
92.8
GCG
Backbone Model=GPT-5.2...
2026.04
68.2
12.5
SkillTrojan
Backbone Model=GLM-4.7
2026.04
67.6
66.4
Non-attack
Backbone Model=GLM-4.7
2026.04
67.3
0
SkillTrojan
Backbone Model=Qwen3-C...
2026.04
65.7
63.9
Non-attack
Backbone Model=Qwen3-C...
2026.04
64.8
0
GCG
Backbone Model=Qwen3-C...
2026.04
62.1
10.3
GCG
Backbone Model=GLM-4.7
2026.04
55.9
26.1
Feedback
Search any
task
Search any
task