List your company
ResearchFormerly WizWand

Vision-Language Reasoning on DisasterBench (val)

87.7Overall AccuracyGemini-2.5-pro
48.408858.609468.8179.0106Jun 4, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Gemini-2.5-proModel Source Type=Closed-Source Models2026.0687.71,323
Claude-Sonnet-4-5Model Source Type=Closed-Source Models2026.0686.26457
GPT-5Model Source Type=Closed-Source Models2026.0683.69644
Gemini-2.5-flashModel Source Type=Closed-Source Models2026.0682.961,321
Gemini-2.0-flashModel Source Type=Closed-Source Models2026.0681.68435
DisasterVLModel Source Type=Open-Source Models, Scale=2B2026.0681.57168
Grok-4Model Source Type=Closed-Source Models2026.0677.87733
Qwen2.5-VL-7BModel Source Type=Open-Source Models2026.0674.94425
GPT-4oModel Source Type=Closed-Source Models2026.0673.19335
InternVL-3-5-4BModel Source Type=Open-Source Models2026.0672.41459
InternVL-3-5-8BModel Source Type=Open-Source Models2026.0672.31322
InternVL2-8BModel Source Type=Open-Source Models2026.0671.38255
Deepseek-VL2-SModel Source Type=Open-Source Models2026.0670.97359
GPT-4o miniModel Source Type=Closed-Source Models2026.0667.27423
Phi-3-Vision(4.15B)Model Source Type=Open-Source Models2026.0666.8272
InternVL-3-5-2BModel Source Type=Open-Source Models2026.0666.13355
Qwen2.5-VL-3BModel Source Type=Open-Source Models2026.0665.31428
InternVL-3-5-1BModel Source Type=Open-Source Models2026.0655.79380
Janus-Pro-1BModel Source Type=Open-Source Models2026.0655.12354
InternVL2-2BModel Source Type=Open-Source Models2026.0654.09425
Deepseek-VL2-TModel Source Type=Open-Source Models2026.0652.91190
Qwen2-VL-2BModel Source Type=Open-Source Models2026.0649.92352