Phishing Detection on Infrastructure Phishing infra_aware
98.6Infra RecallGPT-5.2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5.2Prompt Configuration=infra_aware, Tier=Overcorrected2026.03 | 98.6 | 50.7 | 36.1 | |
| Claude Sonnet 4.5Prompt Configuration=infra_aware, Tier=Overcorrected2026.03 | 98.6 | 32.8 | 51.9 | |
| DeepSeek v3.2Prompt Configuration=infra_aware, Tier=Overcorrected2026.03 | 98.6 | 20.5 | 79.9 | |
| Claude Haiku 4.5Prompt Configuration=infra_aware, Tier=Overcorrected2026.03 | 95.9 | 32.9 | 67.6 | |
| Mistral SmallPrompt Configuration=infra_aware, Tier=Overcorrected2026.03 | 94.5 | 58.9 | 51.1 | |
| Qwen 3Prompt Configuration=infra_aware, Tier=Degraded2026.03 | 91.8 | 2.7 | 38.8 | |
| Gemini 3 FlashPrompt Configuration=infra_aware, Tier=Moderate2026.03 | 90.4 | 76.7 | 20.9 | |
| Grok 4.1 FastPrompt Configuration=infra_aware, Tier=Moderate2026.03 | 84.9 | 65.7 | 15.9 | |
| GPT-4o-miniPrompt Configuration=infra_aware, Tier=Efficient2026.03 | 76.7 | 46.6 | 8.2 | |
| Llama 4 ScoutPrompt Configuration=infra_aware, Tier=Efficient2026.03 | 65.8 | 46.6 | 7.4 | |
| Gemini 2.5 FlashPrompt Configuration=infra_aware, Tier=Efficient2026.03 | 57.5 | 50.7 | 2.8 |