DeepSeek DeepSeek-V3.2 Thinking vs OpenAI GPT-4.1
Detailed comparison for LLMs
DeepSeekOpenAI
Head-to-Head Overview
GPT-4.1 is the overall winner in this comparison!
LLM Metric
Comparison of Attack Success Rate (ASR) metrics for DeepSeek DeepSeek-V3.2 Thinking and OpenAI GPT-4.1 across the three attack methods.
Overall (ASR)
DeepSeek-V3.2 Thinking
100.000
GPT-4.1
34.900
TAP Attack Method (ASR)
DeepSeek-V3.2 Thinking
100.000
GPT-4.1
57.580
Crescendo Attack Method (ASR)
DeepSeek-V3.2 Thinking
100.000
GPT-4.1
44.850
Zero-Shot (ASR)
DeepSeek-V3.2 Thinking
100.000
GPT-4.1
2.290
Key Highlights
- OpenAI GPT-4.1 has a lower Overall (ASR).
- OpenAI GPT-4.1 has a lower TAP Attack Method (ASR).
- OpenAI GPT-4.1 has a lower Crescendo Attack Method (ASR).
- OpenAI GPT-4.1 has a lower Zero-Shot (ASR).
Metric Radar Chart
deepseek-v3-2-thinking
gpt-4-1