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

Web-search QA on BrowseComp-VL

54.9Pass@1

GPT-5

Updated 27d ago

Evaluation Results

Method	Links
GPT-5 2026.05		54.9
GPT-5 2026.05		48.6
OpenSearch-VL-32B 2026.05		43.8
Gemini-2.5-Pro 2026.05		43.1
OpenSearch-VL-30B-A3B 2026.05		41.1
OpenSearch-VL-8B 2026.05		37.6
Gemini-2.5-Flash 2026.05		37.1
SenseNova-MARS-8B 2026.05		35.1
Qwen3-VL-32B 2026.05		35.1
Qwen3-VL-30B-A3B 2026.05		34.1
Claude-3.7-Sonnet 2026.05		32.3
Qwen3-VL-32B 2026.05		30.8
Qwen3-VL-30B-A3B 2026.05		29.6
Claude-4-Sonnet 2026.05		29.3
Qwen3-VL-8B 2026.05		29.3
Qwen3-VL-8B 2026.05		27.9
WebWatcher-32B 2026.05		27
Qwen3-VL-8B 2026.05		24.1
WebWatcher-7B 2026.05		21.2
MMSearch-R1-7B 2026.05		20.9
Visual-ARFT-7B 2026.05		16.5
GPT-4o 2026.05		13.4
Claude-3.7-Sonnet 2026.05		10
GPT-4o 2026.05		5.5