Visual Grounding on RefCOCOg (test)
93.3AccuracyVanilla
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VanillaBackbone Model=DeepSeek-VL2 Small-16B, Average Tokens=100%2026.02 | 93.3 | — | — | — | — | |
| Youtu-VLModel Size=4B2026.01 | 92.9 | — | — | — | — | |
| IVC-PruneBackbone Model=DeepSeek-VL2 Small-16B, Average Tokens=52%2026.02 | 92.4 | — | — | — | — | |
| InternVL2.5Number of Parameters=78B2025.02 | 92.2 | — | — | — | — | |
| Florence-2Backbone=DaViT-B2026.01 | 91.7 | — | — | — | — | |
| Florence-2-L#PT=126M, #FT=127K2025.03 | 91.7 | — | — | — | — | |
| PDropBackbone Model=DeepSeek-VL2 Small-16B, Average Tokens=57%2026.02 | 91.5 | — | — | — | — | |
| InternVL-3-78BBackbone=InternVL-3-78B2026.01 | 91.5 | — | — | — | — | |
| W-FastVBackbone Model=DeepSeek-VL2 Small-16B, Average Tokens=54%2026.02 | 91.3 | — | — | — | — | |
| Qwen3-VL-8B-Thinking + EGMBackbone=Qwen3-VL-8B, Enhancement=EGM2026.01 | 91.2 | — | — | — | — | |
| Qwen3-VL-4B-Thinking + EGMBackbone=Qwen3-VL-4B, Enhancement=EGM2026.01 | 90.8 | — | — | — | — | |
| Qwen3-VL-235B-A22B-InstructBackbone=Qwen3-VL-235B-A22B, Variant=Instruct2026.01 | 90.8 | — | — | — | — | |
| CogVLM-G2026.05 | 90.8 | — | — | — | — | |
| CogVLM-Grounding-17BModel Size=17B, Zero-shot=false2024.02 | 90.79 | — | — | — | — | |
| InternVL-3-8B + EGMBackbone=InternVL-3-8B, Enhancement=EGM2026.01 | 90.7 | — | — | — | — | |
| FastVBackbone Model=DeepSeek-VL2 Small-16B, Average Tokens=54%2026.02 | 90.6 | — | — | — | — | |
| Qwen3-VL-235B-A22B-ThinkingBackbone=Qwen3-VL-235B-A22B, Variant=Thinking2026.01 | 90.5 | — | — | — | — | |
| Qwen2.5-VLNumber of Parameters=72B2025.02 | 90.3 | — | — | — | — | |
| Original Grounding CogVLM2024.06 | 90.2 | — | — | 0 | — | |
| InternVL-3-8BBackbone=InternVL-3-8B2026.01 | 90 | — | — | — | — | |
| InfMLLM-7BLLM=Vicuna-7B2023.11 | 89.78 | — | — | — | — | |
| UNINEXTModel Category=Specialist2024.03 | 89.73 | — | — | — | — | |
| Qwen2.5-VL-7B AIFAIF=true2026.04 | 89.7 | — | — | — | — | |
| GR3D-8B2026.05 | 89.7 | — | — | — | — | |
| UNINEXT#PT=600K, #FT=127K2025.03 | 89.4 | — | — | — | — | |
| UNINEXT-H2026.05 | 89.4 | — | — | — | — | |
| InternVL3.5-8B2026.05 | 89.4 | — | — | — | — | |
| InternVL-3.5Model Size=4B2026.01 | 89.3 | — | — | — | — | |
| Qwen3-VL-2B-Thinking + EGMBackbone=Qwen3-VL-2B, Enhancement=EGM2026.01 | 89.3 | — | — | — | — | |
| Ferretv2#PT=1.1M, #FT=127K2025.03 | 89.3 | — | — | — | — | |
| ONE-PEACE2026.05 | 89.3 | — | — | — | — | |
| UFOBackbone=InternVL2.5-8B2026.01 | 89.2 | — | — | — | — | |
| Non-merging CogVLM2024.06 | 88.9 | — | — | 0 | — | |
| Qwen3-VL-8B-InstructBackbone=Qwen3-VL-8B, Variant=Instruct2026.01 | 88.7 | — | — | — | — | |
| InternVL-3-2B + EGMBackbone=InternVL-3-2B, Enhancement=EGM2026.01 | 88.7 | — | — | — | — | |
| Qwen3-VL-8B-ThinkingBackbone=Qwen3-VL-8B, Variant=Thinking2026.01 | 88.6 | — | — | — | — | |
| TextHawk22026.05 | 88.1 | — | — | — | — | |
| Phi3V-dist#PT=0, #FT=127K2025.03 | 88 | — | — | — | — | |
| Qwen3-VL-4B-InstructBackbone=Qwen3-VL-4B, Variant=Instruct2026.01 | 87.8 | — | — | — | — | |
| HPE-CogVLM2024.06 | 87.8 | — | — | 0 | — | |
| Qwen2-VL-7B2026.05 | 87.8 | — | — | — | — | |
| Qwen3-VLModel Size=4B2026.01 | 87.7 | — | — | — | — | |
| Qwen3-VL-4B-ThinkingBackbone=Qwen3-VL-4B, Variant=Thinking2026.01 | 87.7 | — | — | — | — | |
| VanillaBackbone Model=InternVL 2.5 8B, Average Tokens=100%2026.02 | 87.6 | — | — | — | — | |
| InternVL-2.5-8B2026.04 | 87.6 | — | — | — | — | |
| Phi3V-vocab#PT=0, #FT=127K2025.03 | 87.6 | — | — | — | — | |
| InternVL-3-1B + EGMBackbone=InternVL-3-1B, Enhancement=EGM2026.01 | 87.5 | — | — | — | — | |
| Phi3V-sft#PT=0, #FT=127K2025.03 | 87.4 | — | — | — | — | |
| VanillaBackbone=QWEN-2.5-VL 7B, Average Tokens=100%2026.02 | 87.24 | — | — | — | — | |
| Qwen2.5-VLNumber of Parameters=7B2025.02 | 87.2 | — | — | — | — | |
| GriffonBackbone=LLama2-13B2026.01 | 87.2 | — | — | — | — | |
| InternVL-3-2BBackbone=InternVL-3-2B2026.01 | 87.2 | — | — | — | — | |
| Qwen2.5-VL-7B2026.04 | 87.2 | — | — | — | — | |
| InternVL3-2B2026.05 | 87.2 | — | — | — | — | |
| Qwen2.5-VL-7B2026.05 | 87.2 | — | — | — | — | |
| G-DINO-LModel Category=Specialist2024.03 | 87.02 | — | — | — | — | |
| Grounding DINO2025.02 | 87 | — | — | — | — | |
| Grounding DINOBackbone=Swin-L2026.01 | 87 | — | — | — | — | |
| Grounding-DINO-L2026.04 | 87 | — | — | — | — | |
| Grounding-DINO-L2026.05 | 87 | — | — | — | — | |
| VanillaBackbone Model=Qwen2.5-VL 7B, Average Tokens=100%2026.02 | 86.9 | — | — | — | — | |
| IVC-PruneBackbone Model=InternVL 2.5 8B, Average Tokens=50%2026.02 | 86.9 | — | — | — | — | |
| NuwaBackbone=QWEN-2.5-VL 7B, Average Tokens=75%2026.02 | 86.87 | — | — | — | — | |
| IVC-PruneBackbone Model=Qwen2.5-VL 7B, Average Tokens=50%2026.02 | 86.5 | — | — | — | — | |
| Ferret#PT=1.1M, #FT=127K2025.03 | 86.3 | — | — | — | — | |
| TA merging CogVLMlambda=0.52024.06 | 86.1 | — | — | 0 | — | |
| Qwen2.5-VLNumber of Parameters=3B2025.02 | 85.7 | — | — | — | — | |
| Qwen3-VL-2B-InstructBackbone=Qwen3-VL-2B, Variant=Instruct2026.01 | 85.7 | — | — | — | — | |
| Qwen2.5-VL-3B2026.05 | 85.7 | — | — | — | — | |
| Localization Heads2026.04 | 85.5 | — | — | — | — | |
| InternVL3.5-2B2026.05 | 85.5 | — | — | — | — | |
| NuwaBackbone=QWEN-2.5-VL 7B, Average Tokens=50%2026.02 | 85.49 | — | — | — | — | |
| QWen-VL-7BLLM=Qwen-7B2023.11 | 85.48 | — | — | — | — | |
| Qwen-VL-7BModel Size=7B, Zero-shot=false2024.02 | 85.48 | — | — | — | — | |
| STORM2026.04 | 85.1 | — | — | — | — | |
| Ferret2026.04 | 84.8 | — | — | — | — | |
| Ferret-7BModel Size=7B, Zero-shot=false2024.02 | 84.76 | — | — | — | — | |
| Individual VQA2025.05 | 84.72 | — | — | — | — | |
| M-GPT22026.04 | 84.7 | — | — | — | — | |
| OptMerge2025.05 | 84.68 | — | — | — | — | |
| MiniGPT-v2-7BLLM=LLaMA2-7B2023.11 | 84.66 | — | — | — | — | |
| Individual OCR2025.05 | 84.62 | — | — | — | — | |
| Individual Geometry2025.05 | 84.56 | — | — | — | — | |
| Qwen3-VL-2B-ThinkingBackbone=Qwen3-VL-2B, Variant=Thinking2026.01 | 84.5 | — | — | — | — | |
| Lumen-7BModel Category=Generalist2024.03 | 84.44 | — | — | — | — | |
| VistaLLM-7BModel Size=7B, Zero-shot=false2024.02 | 84.4 | — | — | — | — | |
| Individual Chart2025.05 | 84.33 | — | — | — | — | |
| VisionLLM v2Backbone=Swin-T2026.01 | 84.1 | — | — | — | — | |
| Qwen2.5-VL-32B-Instruct2025.05 | 84.06 | — | — | — | — | |
| Individual Grounding2025.05 | 84.04 | — | — | — | — | |
| VScanBackbone Model=Qwen2.5-VL 7B, Average Tokens=50%2026.02 | 83.6 | — | — | — | — | |
| MDETRBackbone=ENB32026.01 | 83.3 | — | — | — | — | |
| Shik-13BParameters=13B2026.04 | 83.2 | — | — | — | — | |
| Shikra-13BLLM=Vicuna-13B2023.11 | 83.16 | — | — | — | — | |
| Shikra2026.02 | 83.1 | — | — | — | — | |
| InternVL2-8B2026.04 | 82.7 | — | — | — | — | |
| InternVL-3-1BBackbone=InternVL-3-1B2026.01 | 82.6 | — | — | — | — | |
| InternVL3-1B2026.05 | 82.6 | — | — | — | — | |
| CoLLaVO-7BModel Size=7B, Zero-shot=true2024.02 | 82.33 | — | — | — | — | |
| Shik-7BParameters=7B2026.04 | 82.2 | — | — | — | — |