AI Paper Atlas

Research map / Benchmarks / Kinetics-600

Kinetics-600 leaderboard

30 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on Kinetics-600. Best accuracy: 91.8 (TubeViT-H, CVPR 2023).

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.

Kinetics-600 · accuracy

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

#accuracy ↑MethodPaperTopic
191.8TubeViT-H
ViT-H
Rethinking Video ViTs: Sparse Video Tubes for Joint Image and Video Learning
CVPR 2023 · Table 2
Video segmentation and tracking
290.1UniFormerV2-L 336 (64 frames)
CLIP-ViT-L + K710
UniFormerV2: Unlocking the Potential of Image ViTs for Video Understanding
ICCV 2023 · Table 8
Video segmentation and tracking
389.8EVA
ViT giant
EVA: Exploring the Limits of Masked Visual Representation Learning at Scale
CVPR 2023 · Video action recognition table
Open-set and few-shot detection
489.6MTV-H (WTS)Multiview Transformers for Video Recognition
CVPR 2022 · Table 2(b)
Action recognition
588.3MaskFeat MViT-L↑312, 40×3
MViT-L
Masked Feature Prediction for Self-Supervised Visual Pre-Training
CVPR 2022 · Table: Comparison with previou
Video segmentation and tracking
688.3X-CLIP-L/14 8f
CLIP ViT-L/14
Expanding Language-Image Pretrained Models for General Video Recognition
ECCV 2022 · text
Action recognition
787.9MViTv2-L, 352², 40×3MViTv2: Improved Multiscale Vision Transformers for Classification and Detection
CVPR 2022 · Table 7
CNN features
886.3TokenLearner 16at18 w. Fuser (L/10)
ViViT-L/10
TokenLearner: Adaptive Space-Time Tokenization for Videos
NeurIPS 2021 · Table 3
Video segmentation and tracking
986.1Swin-L (384↑)Video Swin Transformer
CVPR 2022 · Table: Comparison to state-of-
Action recognition
1085.6MoViNet-A3-GRW
MoViNet-A3
Smooth Regularization for Efficient Video Recognition
NeurIPS 2025 · Table K600 by FLOPS
Action recognition
20 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.