Model Vulnerability Evaluation on TRAP Text 1.0
91Task Performance ScoreInternVL3.5-14B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| InternVL3.5-14BModel Type=Open-source2026.06 | 91 | 13 | |
| GPT-5-miniModel Type=Proprietary2026.06 | 90 | 2 | |
| GPT-5Model Type=Proprietary2026.06 | 88 | 9 | |
| Qwen3-VL-32BModel Type=Open-source2026.06 | 88 | 1 | |
| Qwen3-VL-8BModel Type=Open-source2026.06 | 87 | 1 | |
| InternVL3.5-8BModel Type=Open-source2026.06 | 87 | 4 | |
| GPT-5.4-miniModel Type=Proprietary2026.06 | 86 | 1 | |
| GPT-4o-miniModel Type=Proprietary2026.06 | 85 | 3 | |
| Llama3.2-11B-VisionModel Type=Open-source2026.06 | 84 | 22 | |
| Gemini-2.5-Flash-LiteModel Type=Proprietary2026.06 | 83 | 13 | |
| Gemini-2.5-FlashModel Type=Proprietary2026.06 | 82 | 2 | |
| InternVL3.5-38BModel Type=Open-source2026.06 | 82 | 7 | |
| Claude Haiku 4.5Model Type=Proprietary2026.06 | 81 | 7 | |
| Devstral-Small-2512Model Type=Open-source2026.06 | 81 | 18 | |
| Claude Sonnet 4.5Model Type=Proprietary2026.06 | 80 | 1 | |
| Qwen3-VL-4BModel Type=Open-source2026.06 | 79 | 16 | |
| Gemini-2.5-ProModel Type=Proprietary2026.06 | 78 | 11 | |
| Phi-4-multimodalModel Type=Open-source2026.06 | 74 | 7 | |
| InternVL3.5-2BModel Type=Open-source2026.06 | 71 | 19 | |
| Qwen3-VL-2BModel Type=Open-source2026.06 | 65 | 10 | |
| InternVL3.5-4BModel Type=Open-source2026.06 | 63 | 9 | |
| GLM-4.6V-FlashModel Type=Open-source2026.06 | 37 | 8 |