List your company

Knowledge-Based Visual Question Answering on KB-VQA Aggregate

37.5Average ScoreInternVL3.5-38B + LoT
9.10816.47923.8531.221Apr 1, 2026

Evaluation results

MethodLinks
InternVL3.5-38B + LoTBackbone=InternVL3.5-38B, Model Scale=Large, Method=LoT2026.0437.5
Qwen3-VL-8B + LoTBackbone=Qwen3-VL-8B, Model Scale=Medium, Method=LoT2026.0435
InternVL3.5-38BBackbone=InternVL3.5-38B, Model Scale=Large, Method=Zero-shot retrieval-augmented2026.0434.1
InternVL3.5-8B + LoTBackbone=InternVL3.5-8B, Model Scale=Medium, Method=LoT2026.0433.7
Qwen3-VL-4B + LoTBackbone=Qwen3-VL-4B, Model Scale=Small, Method=LoT2026.0431.5
Qwen3-VL-8BBackbone=Qwen3-VL-8B, Model Scale=Medium, Method=Zero-shot retrieval-augmented2026.0431.5
Qwen2.5-VL-32B + LoTBackbone=Qwen2.5-VL-32B, Model Scale=Large, Method=LoT2026.0431.5
Qwen3-VL-4BBackbone=Qwen3-VL-4B, Model Scale=Small, Method=Zero-shot retrieval-augmented2026.0430.2
InternVL3.5-8BBackbone=InternVL3.5-8B, Model Scale=Medium, Method=Zero-shot retrieval-augmented2026.0430.2
InternVL3.5-4B + LoTBackbone=InternVL3.5-4B, Model Scale=Small, Method=LoT2026.0429.8
Qwen2.5-VL-7B + LoTBackbone=Qwen2.5-VL-7B, Model Scale=Medium, Method=LoT2026.0429.1
Qwen2-VL-7B + LoTBackbone=Qwen2-VL-7B, Model Scale=Medium, Method=LoT2026.0428.2
Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B, Model Scale=Medium, Method=Zero-shot retrieval-augmented2026.0428
Qwen2.5-VL-32BBackbone=Qwen2.5-VL-32B, Model Scale=Large, Method=Zero-shot retrieval-augmented2026.0427.8
InternVL3.5-4BBackbone=InternVL3.5-4B, Model Scale=Small, Method=Zero-shot retrieval-augmented2026.0425.6
Qwen2.5-VL-3B + LoTBackbone=Qwen2.5-VL-3B, Model Scale=Small, Method=LoT2026.0425.5
Qwen2-VL-7BBackbone=Qwen2-VL-7B, Model Scale=Medium, Method=Zero-shot retrieval-augmented2026.0422.9
Qwen2.5-VL-3BBackbone=Qwen2.5-VL-3B, Model Scale=Small, Method=Zero-shot retrieval-augmented2026.0421.2
Qwen2-VL-2B + LoTBackbone=Qwen2-VL-2B, Model Scale=Small, Method=LoT2026.0411.9
Qwen2-VL-2BBackbone=Qwen2-VL-2B, Model Scale=Small, Method=Zero-shot retrieval-augmented2026.0410.2