AI Paper Atlas

Research map / Benchmarks / COIN

COIN leaderboard

18 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on COIN. Best accuracy: 96.4 (Zero-Shot-CoT + TSP + ARI, ICML 2025).

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.

COIN · accuracy

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

#accuracy ↑MethodPaperTopic
196.4Zero-Shot-CoT + TSP + ARI
GPT-3.5-turbo-0125
Rethinking Chain-of-Thought from the Perspective of Self-Training
ICML 2025 · Table 1
Reasoning and chain of thought
296HierarQ
Vicuna 7B
HierarQ: Task-Aware Hierarchical Q-Former for Enhanced Video Understanding
CVPR 2025 · Table 2
Video-language understanding
394.53PCMCI
CLIP + TimeSformer
Progressive Cross-Modal Causal Intervention for Long-Term Action Recognition
CVPR 2026 · Table 1
Action recognition
493.5HERMESHERMES: temporal-coHERent long-forM understanding with Episodes and Semantics
ICCV 2025 · Table 2
Video-language understanding
593.3AdaCM2
InstructBLIP Q-Former, Vicuna-7B
AdaCM^2: On Understanding Extremely Long-Term Video with Adaptive Cross-Modality Memory Reduction
CVPR 2025 · Table 2
Video-language understanding
693.2MA-LMM
Vicuna-7B
MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding
CVPR 2024 · Table 2
Video-language understanding
792.72LV-MAE (InternVideo2)LV-MAE: Learning Long Video Representations through Masked-Embedding Autoencoders
ICCV 2025 · Table 2
Video segmentation and tracking
890.42S5Selective Structured State-Spaces for Long-Form Video Understanding
CVPR 2023 · COIN comparison table
Video-language understanding
990.4VideoMamba-M (masked pretrain) f64
VideoMamba-M
VideoMamba: State Space Model for Efficient Video Understanding
ECCV 2024 · Table 5
Video-language understanding
1090MoMa f64
CLIP ViT-L/14
MoMa: Modulating Mamba for Adapting Image Foundation Models to Video Recognition
ICML 2025 · Table 3
Video segmentation and tracking
8 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.