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

Multimodal Search on HR-MMSearch

48.52Accuracy

Gemini-3-Pro

Updated 1mo ago

Evaluation Results

Method	Links
Gemini-3-Pro 2026.06		48.52
GPT-5.2 2026.06		48.2
Gemini-3-Flash 2026.06		41.64
TAPO 2026.06		40.66
SenseNova-MARS 2026.06		40.33
Gemini-2.5-Flash 2026.06		40
Qwen3-VL-8B-Instruct+SAPO 2026.06		38.69
GPT-5 2026.06		38.36
Qwen3-VL-32B-Instruct 2026.06		34.43
Qwen2.5-VL-32B-Instruct 2026.06		33.44
GPT-4o 2026.06		30.16
Qwen3-VL-8B-Instruct 2026.06		27.87
Gemini-3-Pro 2026.06		26.89
GPT-4o-mini 2026.06		26.23
GPT-5.2 2026.06		24.92
Visual-ARFT 2026.06		24.92
GPT-5 2026.06		22.62
Gemini-3-Flash 2026.06		21.97
MMSearch-R1 2026.06		20.33
Qwen2.5-VL-7B-Instruct 2026.06		19.34
Qwen3-VL-32B-Instruct 2026.06		19.02
GPT-4o 2026.06		13.11
Qwen3-VL-8B-Instruct 2026.06		12.13
Gemini-2.5-Flash 2026.06		7.54
Qwen2.5-VL-32B-Instruct 2026.06		3.93
GPT-4o-mini 2026.06		1.31
Qwen2.5-VL-7B-Instruct 2026.06		0.58