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

Web-search QA on VDR

33.8Pass@1

OpenSearch-VL-32B

Updated 2mo ago

Evaluation Results

Method	Links
OpenSearch-VL-32B 2026.05		33.8
OpenSearch-VL-30B-A3B 2026.05		33.5
Qwen3-VL-32B 2026.05		23.1
GPT-5 2026.05		22.3
OpenSearch-VL-8B 2026.05		20.8
Qwen3-VL-30B-A3B 2026.05		20.2
SenseNova-MARS-8B 2026.05		19.4
Qwen3-VL-8B 2026.05		17
Claude-3.7-Sonnet 2026.05		11.3
WebWatcher-7B 2026.05		10.3
GPT-5 2026.05		9.8
Gemini-2.5-Pro 2026.05		8
DeepEyes-v2-7B 2026.05		7.8
Qwen3-VL-8B 2026.05		7.3
Gemini-2.5-Flash 2026.05		6.2
Claude-3.7-Sonnet 2026.05		4.6
GPT-4o 2026.05		4.5
Qwen3-VL-32B 2026.05		4.1
Qwen3-VL-30B-A3B 2026.05		3.8
Visual-ARFT-7B 2026.05		3.3
MMSearch-R1-7B 2026.05		2.9
Qwen3-VL-8B 2026.05		2.8
Claude-4-Sonnet 2026.05		2
GPT-4o 2026.05		1.7