Anthropic Claude 3.5 Sonnet v2 vs DeepSeek DeepSeek-V3.2 Thinking
Detailed comparison for LLMs
AnthropicDeepSeek
Head-to-Head Overview
DeepSeek-V3.2 Thinking is the overall winner in this comparison!
LLM Metric
Comparison of Attack Success Rate (ASR) metrics for Anthropic Claude 3.5 Sonnet v2 and DeepSeek DeepSeek-V3.2 Thinking across the three attack methods.
Overall (ASR)
Claude 3.5 Sonnet v2
4.390
DeepSeek-V3.2 Thinking
0
TAP Attack Method (ASR)
Claude 3.5 Sonnet v2
8.830
DeepSeek-V3.2 Thinking
0
Crescendo Attack Method (ASR)
Claude 3.5 Sonnet v2
3.810
DeepSeek-V3.2 Thinking
0
Zero-Shot (ASR)
Claude 3.5 Sonnet v2
0.540
DeepSeek-V3.2 Thinking
0
Key Highlights
- DeepSeek DeepSeek-V3.2 Thinking has a lower Overall (ASR).
- DeepSeek DeepSeek-V3.2 Thinking has a lower TAP Attack Method (ASR).
- DeepSeek DeepSeek-V3.2 Thinking has a lower Crescendo Attack Method (ASR).
- DeepSeek DeepSeek-V3.2 Thinking has a lower Zero-Shot (ASR).
Metric Radar Chart
claude-3-5-sonnet-v2
deepseek-v3-2-thinking