List your company

Model Vulnerability Evaluation on TRAP Text 1.0

91Task Performance ScoreInternVL3.5-14B
34.8449.426478.58Jun 17, 2026

Evaluation results

MethodLinks
InternVL3.5-14BModel Type=Open-source2026.069113
GPT-5-miniModel Type=Proprietary2026.06902
GPT-5Model Type=Proprietary2026.06889
Qwen3-VL-32BModel Type=Open-source2026.06881
Qwen3-VL-8BModel Type=Open-source2026.06871
InternVL3.5-8BModel Type=Open-source2026.06874
GPT-5.4-miniModel Type=Proprietary2026.06861
GPT-4o-miniModel Type=Proprietary2026.06853
Llama3.2-11B-VisionModel Type=Open-source2026.068422
Gemini-2.5-Flash-LiteModel Type=Proprietary2026.068313
Gemini-2.5-FlashModel Type=Proprietary2026.06822
InternVL3.5-38BModel Type=Open-source2026.06827
Claude Haiku 4.5Model Type=Proprietary2026.06817
Devstral-Small-2512Model Type=Open-source2026.068118
Claude Sonnet 4.5Model Type=Proprietary2026.06801
Qwen3-VL-4BModel Type=Open-source2026.067916
Gemini-2.5-ProModel Type=Proprietary2026.067811
Phi-4-multimodalModel Type=Open-source2026.06747
InternVL3.5-2BModel Type=Open-source2026.067119
Qwen3-VL-2BModel Type=Open-source2026.066510
InternVL3.5-4BModel Type=Open-source2026.06639
GLM-4.6V-FlashModel Type=Open-source2026.06378