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

Document Visual Question Answering on SlideVQA

0.772F1 Score

Doc-V* (GRPO)

Updated 3d ago

Evaluation Results

Method	Links
Doc-V* (GRPO) 2026.04		0.772	-
Claude-3.7-Sonnet 2025.12		0.763	0.629
Claude-3.7-Sonnet 2026.04		0.763	-
GPT-4.1 2025.12		0.747	0.615
GPT-4.1 2026.04		0.747	-
Doc-V* (SFT) 2026.04		0.738	-
CogDoc 2025.12		0.679	0.583
CogDoc 2026.04		0.679	-
GPT-4o 2025.12		0.658	0.531
GPT-4o 2026.04		0.658	-
InternVL3-8B 2025.12		0.644	0.516
InternVL3 2026.04		0.644	-
FRAG 2025.12		0.639	0.524
Qwen2.5-VL (RAG Top-5) 2026.04		0.629	-
MiMo-VL-7B-SFT 2025.12		0.619	0.48
GPT-4o mini 2025.12		0.607	0.481
GPT-4o mini 2026.04		0.607	-
M3DocRAG 2025.12		0.579	0.488
M3DocRAG 2026.04		0.557	-
Qwen2.5-VL (Baseline) 2026.04		0.552	-
Qwen2.5-VL-72B-Instruct 2025.12		0.544	0.388
VisRAG 2025.12		0.524	0.509
VisRAG 2026.04		0.524	-
Qwen2.5-VL-7B-Instruct 2025.12		0.498	0.389
VDocRAG 2025.12		0.44	0.42
Docopilot 2025.12		0.431	0.357
Docopilot 2026.04		0.431	-
VDocRAG 2026.04		0.42	-
SV-RAG 2025.12		0.343	0.32
SV-RAG 2026.04		0.343	-
mPLUG-DocOwl2 2025.12		0.304	0.254
mPLUG-DocOwl2 2026.04		0.278	-
Vanilla RAG 2026.02		-	0.448
ReAct 2026.02		-	0.449
UniversalRAG 2026.02		-	0.147
VideoRAG 2026.02		-	0.34
MemAgent 2026.02		-	0.431
Mem1 2026.02		-	0.498
VimRAG 2026.02		-	0.528
Vanilla RAG 2026.02		-	0.485
ReAct 2026.02		-	0.5
UniversalRAG 2026.02		-	0.172
VideoRAG 2026.02		-	0.355
MemAgent 2026.02		-	0.453
Mem1 2026.02		-	0.557
VimRAG 2026.02		-	0.624