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

Visual Grounding on RefCOCO (val)

95.2Accuracy

Vanilla

Updated 23d ago

Evaluation Results

Method
Vanilla 2026.02	95.2	-	-
Phi3V-dist 2025.03	94.8	-	-
IVC-Prune 2026.02	94.5	-	-
Phi3V-vocab 2025.03	94.5	-	-
Phi3V-sft 2025.03	94.3	-	-
Qwen3-VL-8B-Thinking + EGM 2026.01	93.9	-	-
InternVL2.5 2025.02	93.7	-	-
W-FastV 2026.02	93.6	-	-
Youtu-VL 2026.01	93.6	-	-
Youtu-VL 2026.01	93.6	-	-
InternVL-3-8B + EGM 2026.01	93.6	-	-
Qwen3-VL-4B-Thinking + EGM 2026.01	93.5	-	-
Florence-2 2026.01	93.4	-	-
InternVL-3-78B 2026.01	93.4	-	-
OpenVLThinkerV2 2026.04	93.4	-	-
Florence-2-L 2025.03	93.4	-	-
UFO 2026.01	93.1	-	-
PDrop 2026.02	93	-	-
Qwen3-VL-2B-Thinking + EGM 2026.01	93	-	-
Qwen3-VL-235B-A22B-Instruct 2026.01	92.9	-	-
Qwen3-VL-235B-A22B-Thinking 2026.01	92.9	-	-
Ferretv2 2025.03	92.8	-	-
CogVLM-Grounding-17B 2024.02	92.76	-	-
Qwen2.5-VL 2025.02	92.7	-	-
UNINEXT 2024.03	92.64	-	-
FastV 2026.02	92.6	-	-
Classic Specialist (VLM) 2026.01	92.6	-	-
UNINEXT 2025.03	92.6	-	-
Florence-2-B 2025.03	92.6	-	-
ONE-PEACE 2023.05	92.58	-	-
InternVL-3.5 2026.01	92.5	-	-
InternVL-3.5 2026.01	92.5	-	-
InternVL-3-8B 2026.01	92.5	-	-
InternVL3-8B 2026.02	92.5	-	-
InternVL-3.5 2026.05	92.5	-	-
Qwen3-VL-Seg 2026.05	92.3	-	-
InternVL-3-2B + EGM 2026.01	92.2	-	-
Qwen3-VL 2026.04	92.2	-	-
Qwen3-VL 2026.04	92.1	-	-
OFA 2023.05	92.04	-	-
OneThinker-8B 2026.04	92	-	-
UFO 2026.01	91.8	-	-
Qwen3-VL 2026.05	91.8	-	-
Qwen3-VL-8B-Instruct 2026.01	91.6	-	-
Qwen2.5-VL-7B AIF 2026.04	91.6	-	-
Qwen3-VL 2026.05	91.6	-	-
RoboInter-Qwen-7B 2026.02	91.5	-	-
RoboInter-LLaVAOV-7B 2026.02	91.3	-	-
Qwen3-VL-4B-Instruct 2026.01	91	-	-
Qwen3-VL-8B-Thinking 2026.01	91	-	-
Qwen3-VL 2026.01	90.7	-	-
Qwen3-VL 2026.01	90.7	-	-
FIBER 2023.05	90.68	-	-
Grounding DINO 2025.02	90.6	-	-
Grounding DINO 2026.01	90.6	-	-
Grounding DINO 2026.04	90.6	-	-
Grounding-DINO-L 2026.04	90.6	-	-
Grounding-DINO 2023.05	90.56	-	-
G-DINO-L 2024.03	90.56	-	-
Classic Specialist (Non-VLM) 2026.01	90.5	-	-
VLM-R1 2026.04	90.5	-	-
Vanilla 2026.02	90.3	-	-
IVC-Prune 2026.02	90.3	-	-
InternVL-2.5-8B 2026.04	90.3	-	-
InternVL-3-1B + EGM 2026.01	90.2	-	-
QwenVL2.5-7B 2026.02	90.2	-	-
Griffon 2026.01	90.1	-	-
Qwen2.5-VL 2025.02	90	-	-
VisionLLM v2 2026.01	90	-	-
Qwen3-VL-4B-Thinking 2026.01	90	-	-
Qwen2.5-VL-7B 2026.04	90	-	-
Qwen3-VL 2026.04	89.9	-	-
InternVL-3-2B 2026.01	89.8	-	-
DeepEyes 2026.04	89.8	-	-
Vanilla 2026.02	89.6	-	-
RoboInter-Qwen-3B 2026.02	89.5	-	-
Ferret 2025.03	89.5	-	-
QWen-VL-7B 2023.11	89.36	-	-
Qwen-VL-7B 2024.02	89.36	-	-
IVC-Prune 2026.02	89.3	-	-
Qwen2.5-VL 2025.02	89.1	-	-
QwenVL2.5-3B 2026.02	89.1	-	-
STORM 2026.04	89.1	-	-
Qwen3-VL-2B-Instruct 2026.01	88.7	-	-
M-GPT2 2026.04	88.7	-	-
MiniGPT-v2-7B 2023.11	88.69	-	-
Lumen-7B 2024.03	88.59	-	-
UNICORN 2023.05	88.29	-	-
VistaLLM-7B 2024.02	88.1	-	-
G-GPT 2026.04	88	-	-
VisionLLM v2 2026.01	87.9	-	-
Shikra-13B 2023.11	87.83	-	-
Shikra 2026.02	87.8	-	-
Shik-13B 2026.04	87.8	-	-
MDETR 2026.01	87.5	-	-
Qwen3-VL-2B-Thinking 2026.01	87.5	-	-
Ferret 2026.04	87.5	-	-
Ferret-7B 2024.02	87.49	-	-
CoLLaVO-7B 2024.02	87.34	-	-
Localization Heads 2026.04	87.2	-	-

Showing 100 of 192 rows