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

Question Answering on HiTab (held-in)

58.2Accuracy

Qwen2-VL-7B-Table-R1

Updated 2mo ago

Evaluation Results

Method	Links
Qwen2-VL-7B-Table-R1 2025.09		58.2
Qwen2-VL-7B-GRPO 2025.09		54.7
Qwen2-VL-7B-SFT 2025.09		48.5
Gemini-2.5-Pro 2025.09		46.44
QVQ-72B-Preview 2025.09		45.7
OpenAI-o4-mini 2025.09		44.8
Qwen2-VL-72B 2025.09		44.04
Claude-3.5-Sonnet 2025.09		41.31
GPT-4o 2025.09		39.32
Qwen2-VL-2B-Table-R1 2025.09		37.3
Qwen2-VL-2B-GRPO 2025.09		35.2
Qwen2-VL-7B 2025.09		33.1
Qwen2-VL-2B-SFT 2025.09		31.9
Qwen2-VL-2B 2025.09		22.9
DeepSeek-VL2 4.5B 2025.09		18.89
Table-LLaVA 13B 2025.09		10.85
Table-LLaVA 7B 2025.09		10.09
Monkey 7B 2025.09		6.41
Qwen2-VL-7B-Ins 2025.09		5.33
LLaVA-v1.5 7B 2025.09		2.03
mPLUG-Owl2 7B 2025.09		0.13