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

Document-Oriented Visual Question Answering on DocVQA

94.9Accuracy

Vanilla

Updated 1mo ago

Evaluation Results

Method	Links
Vanilla 2026.02		94.9
IVC-Prune 2026.02		94.4
FastV 2026.02		93.9
PDrop 2026.02		93.8
LLaVA-OneVision 2024.09		93.1
Oryx-1.5 2024.09		92.7
Oryx 2024.09		91.4
Oryx-1.5 2024.09		90.1
Oryx 2024.09		89
Florence-VL 8B 2024.12		84.9
Magma-8B (Full) 2025.02		84.8
FastVLM-1.7B 2024.12		84.2
Magma-8B (Actw/o) 2025.02		84.1
LLaVA-NeXT 2024.09		84
Magma-8B (Fullw/o) 2025.02		83.8
Florence-VL 3B 2024.12		82.1
FlorenceVL-3B 2024.12		82.1
MM1.5-1B 2024.12		81
Magma-8B (SFT) 2025.02		80.4
LLaVA-NeXT+CAL 2024.05		80.1
SmolVLM2-2.2B 2024.12		80
FastVLM-0.6B 2024.12		79.1
LLaVA-NeXT 2024.05		78.4
LLaVA-NeXT 2024.09		78.2
MGM-HD+CAL 2024.05		78
Cambrian 8B 2024.12		77.8
Cambrian-1 2024.09		77.8
MGM-HD 2024.05		77.7
MolmoE A1B-7B 2024.12		77.7
LLaVA-NeXT+CAL 2024.05		77.3
Phi 3.5 Vision 2024.12		75.9
Cambrian-1 2024.09		75.5
LLaVA-NeXT 2024.05		75.1
Mini-Gemini-HD 8B 2024.12		74.6
LLaVA-Next-7B 2025.02		74.4
MGM-HD+CAL 2024.05		73.4
MGM-HD 2024.05		72
SmolVLM2-0.5B 2024.12		70.5
LLaVA next 8B 2024.12		69.3
MGM 2024.05		61.7
MGM+CAL 2024.05		61.6
MGM+CAL 2024.05		58
MGM 2024.05		57.7
Monkey 2024.07		50.1
TextHarmony-Chat 2024.07		49.8
DocPedia 2024.07		47.1
TextHarmony 2024.07		47.1
MGM+CAL 2024.05		44.8
TextHarmony* 2024.07		44
MGM 2024.05		39.8
InternLM-XComposer2 2024.07		39.7
LLaVA-1.5+CAL 2024.05		32.6
LLaVA-1.5 2024.05		31.1
LLaVA-1.5+CAL 2024.05		30.6
AIMv2 2024.11		30.4
InternVL 2024.07		28.7
LLaVA-1.5 2024.05		28.5
AIMv2 2024.11		28.5
LLaVA-1.5-7B 2025.02		28.1
AIMv2 2024.11		27.8
AIMv2 2024.11		26.6
OpenAI CLIP 2024.11		25.6
SigLIP 2024.11		19.2
mPLUG-Owl2 2024.07		17.9
SigLIP 2024.11		16.9
LLaVAR 2024.07		12.3
LLaVA1.5-7B 2024.07		8.5
DINOv2 2024.11		8.2
MM-Interleaved 2024.07		8.1
UniDoc 2024.07		7.7
SEED-LLaMA-14B 2024.07		6.5
MiniGPT5 2024.07		1.6