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

Question Answering on TabMWP (held-in)

92.6Accuracy

Qwen2-VL-7B-Table-R1

Updated 22d ago

Evaluation Results

Method	Links
Qwen2-VL-7B-Table-R1 2025.09		92.6
Qwen2-VL-7B-SFT 2025.09		90.3
Gemini-2.5-Pro 2025.09		89.9
Qwen2-VL-7B-GRPO 2025.09		89.2
GPT-4o 2025.09		87.59
OpenAI-o4-mini 2025.09		86.7
QVQ-72B-Preview 2025.09		86.2
Claude-3.5-Sonnet 2025.09		83.3
Qwen2-VL-2B-Table-R1 2025.09		83.2
Qwen2-VL-72B 2025.09		81.95
Qwen2-VL-2B-GRPO 2025.09		71.4
Qwen2-VL-2B-SFT 2025.09		70
Qwen2-VL-7B 2025.09		63.8
Table-LLaVA 13B 2025.09		59.77
Table-LLaVA 7B 2025.09		57.78
DeepSeek-VL2 4.5B 2025.09		53.75
Qwen2-VL-7B-Ins 2025.09		49.51
Qwen2-VL-2B 2025.09		46.1
Monkey 7B 2025.09		13.26
mPLUG-Owl2 7B 2025.09		6.83
LLaVA-v1.5 7B 2025.09		6.05