Referring Expression Comprehension on RefCOCO, RefCOCO+, and RefCOCOg Average benchmark leaderboard · SOTA2 Research

Referring Expression Comprehension on RefCOCO, RefCOCO+, and RefCOCOg Average

92.3Average AccuracyInternVL2.5-78B

Evaluation Results

Method	Links
InternVL2.5-78B2024.12		92.3
Qwen2-VL-72B2024.12		91.1
CogVLM-Grounding-17B2024.12		90.3
CogVLM-17B2024.07		90.25
InternVL2-Llama3-76B2024.12		90
ONE-PEACE2024.12		89.8
Ferret-v2-13B2024.12		89.6
UNINEXT-H2024.12		88.9
TextHawk22024.12		88.2
Qwen2-VL-7B2024.12		87.9
CoS-7B2024.07		87.68
InternVL2.5-8B2024.12		87.6
UNINEXT-L2024.04		86.95
G-DINO2024.04		86.6
Grounding-DINO-L2024.12		86.6
Groma2024.04		86.52
Qwen-VL2024.04		85.83
Qwen-VL-7B2024.07		85.83
Ferret-13B2024.07		85.57
Griffon v22024.07		85.42
MiniGPT-v22024.04		84.29
MiniGPTv22024.07		84.29
Shikra-13B2024.07		83.95
Ferret2024.04		83.91
Ferret-7B2024.07		83.91
CoLA2026.04		83.4
Shikra2024.04		82.93
Shikra-7B2024.07		82.93
Shikra-7B2024.12		82.9
InternVL2-8B2024.12		82.9
MDETR2024.04		81.81
MDETR2024.07		81.81
SwimVG2026.04		81.8
EEVG2026.04		81.7
HiVG2026.04		80.9
VG-LAW2026.04		79.7
TransVG++2026.04		78.8
MaPPER2026.04		78.7
QRNet2026.04		76.3
UNITER2024.07		76.17
OFA2024.04		72.65
OFA-L2024.07		72.65
TransVG2026.04		71.4
MAttNet2024.07		68.87