List your company
ResearchFormerly WizWand

Visual Question Answering on FVQA

82.82AccuracyStaR-KVQA_Qwen
21.584837.482453.3869.2776Oct 8, 2025Nov 21, 2025Jan 4, 2026Feb 17, 2026Apr 2, 2026May 16, 2026Jun 30, 2026
Updated 4mo ago

Evaluation results

MethodLinks
StaR-KVQA_QwenBackbone=Qwen2.5-VL-7B, Training=Fine-tuned2025.1082.82
StaR-KVQA_GemmaBackbone=Gemma-3-12B, Training=Fine-tuned2025.1081.2
StaR-KVQA_LlamaBackbone=Llama-3.2-11B-Vision, Training=Fine-tuned2025.1080.19
SimpleSearch-VL-30B-A3BEvaluation Paradigm=Multimodal Search Agents2026.0679
LLaVA-CoTBackbone=Llama-3.2-11B-Vision, Training=Fine-tuned2025.1078.45
Gemma-3-27BBackbone=Gemma-3-27B2025.1076.82
SimpleSearch-VL-8BEvaluation Paradigm=Multimodal Search Agents2026.0676.8
Gemini-3-ProEvaluation Paradigm=Agentic Workflow2026.0676.7
Kimi-K2.5Evaluation Paradigm=Agentic Workflow2026.0676.5
Qwen2.5-VL-72BBackbone=Qwen2.5-VL-72B2025.1075.95
SDFTBackbone=Qwen2.5-VL-7B, Training=Self-Distillation Fine-Tuning2025.1075.54
CoT + SFTBackbone=Qwen2.5-VL-7B, Training=Fine-tuned, Mode=Chain-of-Thought2025.1075.13
CoTBackbone=Qwen2.5-VL-7B, Mode=Chain-of-Thought2025.1074.66
Gemini 2.5 Flash2025.1074.51
Claude-Opus-4.6Evaluation Paradigm=Agentic Workflow2026.0674.5
Vision-DeepResearch-30B-A3BEvaluation Paradigm=Multimodal Search Agents2026.0674.2
SFTBackbone=Qwen2.5-VL-7B, Training=Fine-tuned2025.1073.91
Gemini 2.5 Pro2025.1073.39
OpenSearch-VL-30B-A3BEvaluation Paradigm=Multimodal Search Agents2026.0673.2
SenseNova-MARS-32BEvaluation Paradigm=Multimodal Search Agents2026.0672.6
M2-ReasoningBackbone=M2-Reasoning-7B2025.1072.53
GPT-4o2025.1072.36
Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2025.1071.61
OpenSearch-VL-8BEvaluation Paradigm=Multimodal Search Agents2026.0671.5
POINTS-Seeker-8BEvaluation Paradigm=Multimodal Search Agents2026.0671.2
InternVL3-78BBackbone=InternVL3-78B2025.1070.99
Gemma-3-12BBackbone=Gemma-3-12B2025.1070.64
MM-DeepResearch-8BEvaluation Paradigm=Multimodal Search Agents2026.0669.2
GoG-3-8B-Think-RLEvaluation Methodology=Search-Equipped Models2026.0168.44
Qwen3-VL-30B-A3BEvaluation Paradigm=Agentic Workflow2026.0667.4
SenseNova-MARS-8BEvaluation Paradigm=Multimodal Search Agents2026.0667.1
GoG-2.5-7B-RLEvaluation Methodology=Search-Equipped Models2026.0166.78
Llama-3.2-11B-VisionBackbone=Llama-3.2-11B-Vision2025.1066.09
Qwen3-VL-8BEvaluation Paradigm=Agentic Workflow2026.0665.3
Vision-DeepResearch-8BEvaluation Paradigm=Multimodal Search Agents2026.0664.7
WebWatcher-32BEvaluation Paradigm=Multimodal Search Agents2026.0664.3
GoG-3-8B-Think-SFTEvaluation Methodology=Search-Equipped Models2026.0162.17
Qwen2.5-VL-7BEvaluation Methodology=Full-Search Workflow2026.0161.06
DeepEyesV2-7BEvaluation Paradigm=Multimodal Search Agents2026.0660.6
Claude-Opus-4.6Evaluation Paradigm=Direct Answer2026.0660.1
Kimi-K2.5Evaluation Paradigm=Direct Answer2026.0659.6
Gemini-3-ProEvaluation Paradigm=Direct Answer2026.0658.9
MMSearch-R1-7BEvaluation Paradigm=Multimodal Search Agents2026.0658.4
Qwen3-VL-8B-ThinkingEvaluation Methodology=Full-Search Workflow2026.0157.33
GoG-2.5-7B-SFTEvaluation Methodology=Search-Equipped Models2026.0153.72
Qwen3-VL-8B-ThinkEvaluation Methodology=Prompt-based GoG Method2026.0151.33
MMSearch-R1*Evaluation Methodology=Search-Equipped Models2026.0142.39
GPT-4oEvaluation Methodology=Direct Answer2026.0142
Qwen3-VL-30B-A3BEvaluation Paradigm=Direct Answer2026.0634.7
Qwen3-VL-30B-A3B-ThinkingEvaluation Methodology=Direct Answer2026.0131.11
Qwen3-VL-8BEvaluation Paradigm=Direct Answer2026.0628
Qwen2.5-VL-32BEvaluation Methodology=Direct Answer2026.0126.89
Qwen3-VL-30B-A3B-InstructEvaluation Methodology=Direct Answer2026.0126.11
Qwen2.5-VL-7BEvaluation Methodology=Direct Answer2026.0125.33
Qwen3-VL-8B-ThinkingEvaluation Methodology=Direct Answer2026.0124.56
Qwen2.5-VL-7BEvaluation Methodology=Prompt-based GoG Method2026.0124.5
Qwen3-VL-8B-InstructEvaluation Methodology=Direct Answer2026.0123.94