Share your thoughts, 1 month free Claude Pro on usSee more

Multimodal Question Answering on MMBench en (test)

89Accuracy

Vanilla

Updated 11d ago

Evaluation Results

Method	Links
Vanilla 2025.12		89	100	100
DART (ViT) 2025.12		85	88.1	63.9
InternVL2.5-8B 2026.04		84.6	-	-
IPCV+FastV 2025.12		84.5	89.9	64.4
ToFu 2025.12		83.9	83.4	61.9
GPT-4o 2026.04		83.4	-	-
Qwen2.5-VL-7B 2026.04		82.8	-	-
Claude-3.5 Sonnet 2026.04		82.6	-	-
Insight-V-8B 2026.04		82.3	-	-
ToMe 2025.12		82.2	82.7	61.7
IXC-2.5 2026.04		82.2	-	-
Qwen2.5-VL-7B + Saliency-R1 2026.04		81.8	-	-
InternVL2-8B 2026.04		81.7	-	-
Qwen2.5-VL-7B + SFT 2026.04		78.8	-	-
MiniCPM-V-2.6-8B 2026.04		78	-	-
Qwen2.5-VL-3B 2026.04		77.8	-	-
MiniCPM-Llama-V-2.5-8B 2026.04		77.2	-	-
Qwen2.5-VL-3B + Saliency-R1 2026.04		76.7	-	-
FastV 2025.12		76.6	78.6	70.7
Cambrain-1-8B 2026.04		75.9	-	-
Qwen2.5-VL-3B + SFT 2026.04		75.8	-	-
LLaVA-CoT-11B 2026.04		75	-	-
SparseVLM 2025.12		74.6	75.1	73.4
Gemini-1.5-Pro 2026.04		73.9	-	-
DART 2025.12		71.1	73.6	71.5
V^2Drop 2025.12		59.5	57.1	70.5