Agentic Search on BC-VL
48.6AccuracyClaude-4-Sonnet
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude-4-SonnetWorkflow=Agent Workflow2026.04 | 48.6 | |
| GPT-5Workflow=Direct Answer2026.04 | 47.2 | |
| GPT-5Workflow=Agent Workflow2026.04 | 46.1 | |
| Gemini-2.5 FlashWorkflow=Agent Workflow2026.04 | 44.6 | |
| POINTS-Seeker-8BWorkflow=Agentic Search MLLM2026.04 | 44.4 | |
| Vision-DeepResearch-8BWorkflow=Agentic Search MLLM2026.04 | 42.6 | |
| Skywork-R1V4-30B-A3BWorkflow=Agentic Search MLLM2026.04 | 38.4 | |
| MM-DeepResearch-8BWorkflow=Agentic Search MLLM2026.04 | 37.9 | |
| Gemini-2.5 FlashWorkflow=Direct Answer2026.04 | 37.1 | |
| Qwen3-VL-8B-ThinkingWorkflow=Agent Workflow2026.04 | 37.1 | |
| Claude-4-SonnetWorkflow=Direct Answer2026.04 | 29.3 | |
| WebWatcher-32BWorkflow=Agentic Search MLLM2026.04 | 27 | |
| Qwen3-VL-8B-InstructWorkflow=Direct Answer2026.04 | 25.1 | |
| WebWatcher-7BWorkflow=Agentic Search MLLM2026.04 | 21.2 | |
| GPT-4oWorkflow=RAG Workflow2026.04 | 13.4 | |
| Gemini-2.5 FlashWorkflow=RAG Workflow2026.04 | 13 | |
| Qwen-2.5-VL-72BWorkflow=RAG Workflow2026.04 | 11.5 | |
| Claude-3.7-SonnetWorkflow=RAG Workflow2026.04 | 11.2 |