Visual Document Understanding and Question Answering: A Multi-Agent Collaboration Framework with Test-Time Scaling

Benchmark Model Rank Results
infographic-visual-question-answering-on-infovqaMACT Qwen2.5-VL Series 24B (Agent-Wise Adaptive TTS) (extra data)#2ANLS: 89.4
visual-question-answering-on-docvqa-testMACT Qwen2.5-VL Series 24B (Agent-Wise Adaptive TTS) (extra data)#2ANLS: 0.966