AI Paper Atlas

Research map / Benchmarks / gRefCOCO

gRefCOCO leaderboard

20 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on gRefCOCO. Best gIoU (val): 78.4 (SetCon, NeurIPS 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

gRefCOCO · gIoU (val)

14 papers, each with its best reported number for its own method; higher is better.

#gIoU ↑MethodPaperTopic
178.4SetCon
Qwen3-VL-8B
SetCon: Towards Open-Ended Referring Segmentation via Set-Level Concept Prediction
NeurIPS 2026 · Table 1
Open-set and few-shot detection
278.1ALToLLM-8B (all components)
InternVL-2.5-8B
ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation
NeurIPS 2025 · Table: Ablation study for ALTo
Multimodal LLMs
377.67DeRIS-LDeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy
ICCV 2025 · Table 2
Open-set and few-shot detection
476.8SegCompass-13BSegCompass: Exploring Interpretable Alignment with Sparse Autoencoders for Enhanced Reasoning Segmentation
CVPR 2026 · Table 2
Multimodal LLMs
574.7ARGenSeg
InternVL2.5-8B
ARGenSeg: Image Segmentation with Autoregressive Image Generation Model
NeurIPS 2025 · Table 2
Multimodal LLMs
674.64RAS 13B Co-DETR GRES-FTRefer to Any Segmentation Mask Group With Vision-Language Prompts
ICCV 2025 · Table 2
Open-set and few-shot detection
772.46UGround-LLaVA1.5-7B (ft)
LLaVA1.5-7B
UGround: Towards Unified Visual Grounding with Unrolled Transformers
ICML 2026 · Table 4
Multimodal LLMs
871.86SAM4MLLM-8B-PQPP-LLaVA1.6
LLaVA1.6-8B
SAM4MLLM: Enhance Multi-Modal Large Language Model for Referring Expression Segmentation
ECCV 2024 · Table 3
Human and embodied modeling
971.32Ours
Swin-B
Improving Target Presence and Plurality Recognition for Generalized Referring Image Segmentation
AAAI 2026 · Table 1
Open-set and few-shot detection
1068.42CoHD
Swin-B
CoHD: A Counting-Aware Hierarchical Decoding Framework for Generalized Referring Expression Segmentation
ICCV 2025 · Table 1
Open-set and few-shot detection
4 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

gRefCOCO · cIoU (val)

12 papers, each with its best reported number for its own method; higher is better.

#cIoU ↑MethodPaperTopic
175.5ALToLLM-8B (all components)
InternVL-2.5-8B
ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation
NeurIPS 2025 · Table: Ablation study for ALTo
Multimodal LLMs
272.2SegCompass-13BSegCompass: Exploring Interpretable Alignment with Sparse Autoencoders for Enhanced Reasoning Segmentation
CVPR 2026 · Table 2
Multimodal LLMs
372.2ARGenSeg
InternVL2.5-8B
ARGenSeg: Image Segmentation with Autoregressive Image Generation Model
NeurIPS 2025 · Table 2
Multimodal LLMs
472SetCon
Qwen3-VL-8B
SetCon: Towards Open-Ended Referring Segmentation via Set-Level Concept Prediction
NeurIPS 2026 · Table 1
Open-set and few-shot detection
572DeRIS-LDeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy
ICCV 2025 · Table 2
Open-set and few-shot detection
667.73Segment Anyword_fSegment Anyword: Mask Prompt Inversion for Open-Set Grounded Segmentation
ICML 2025 · gRefCOCO results table
Open-set and few-shot detection
767.36InstructDiffusion
Stable Diffusion v1.5
InstructDiffusion: A Generalist Modeling Interface for Vision Tasks
CVPR 2024 · Table 3
Human and embodied modeling
866.61Ours
Swin-B
Improving Target Presence and Plurality Recognition for Generalized Referring Image Segmentation
AAAI 2026 · Table 1
Open-set and few-shot detection
965.56UGround-LLaVA1.5-7B (ft)
LLaVA1.5-7B
UGround: Towards Unified Visual Grounding with Unrolled Transformers
ICML 2026 · Table 4
Multimodal LLMs
1065.17CoHD
Swin-B
CoHD: A Counting-Aware Hierarchical Decoding Framework for Generalized Referring Expression Segmentation
ICCV 2025 · Table 1
Open-set and few-shot detection
2 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

gRefCOCO · N-acc (val)

12 papers, each with its best reported number for its own method; higher is better.

#N-acc ↑MethodPaperTopic
182.22DeRIS-LDeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy
ICCV 2025 · Table 2
Open-set and few-shot detection
281.3VBackChecker (Ours)Seeing Is Believing: Rich-Context Hallucination Detection for MLLMs via Backward Visual Grounding
AAAI 2026 · Table 2
Multimodal LLMs
374.53UGround-LLaVA1.5-7B (ft)
LLaVA1.5-7B
UGround: Towards Unified Visual Grounding with Unrolled Transformers
ICML 2026 · Table 4
Multimodal LLMs
472.27Ours
Swin-B
Improving Target Presence and Plurality Recognition for Generalized Referring Image Segmentation
AAAI 2026 · Table 1
Open-set and few-shot detection
569.05RAS 13B Co-DETR GRES-FTRefer to Any Segmentation Mask Group With Vision-Language Prompts
ICCV 2025 · Table 2
Open-set and few-shot detection
663.68CoHD
Swin-B
CoHD: A Counting-Aware Hierarchical Decoding Framework for Generalized Referring Expression Segmentation
ICCV 2025 · Table 1
Open-set and few-shot detection
760.3HieA2G
ResNet101
Hierarchical Alignment-enhanced Adaptive Grounding Network for Generalized Referring Expression Comprehension
AAAI 2025 · GREC results table
Human and embodied modeling
859.16LaConvNet-STowards Language-Guided Visual Recognition via Dynamic Convolutions
IJCV 2023 · Table 6
Multimodal LLMs
957.1ViSurf
Qwen2.5-VL-7B
ViSurf: Visual Supervised-and-Reinforcement Fine-Tuning for Large Vision-and-Language Models
ICML 2026 · Table 1
Multimodal LLMs
1056.45GSVA-Vicuna-7B
Vicuna-7B
GSVA: Generalized Segmentation via Multimodal Large Language Models
CVPR 2024 · Table 1
Multimodal LLMs
2 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.