| Universal Video Temporal Grounding with Generative Multi-modal Large Language Models | | 74.33 | 53.71 | | | | | UniTime-SP | 2025-06-23 |
| TempSamp-R1: Effective Temporal Sampling with Reinforcement Fine-Tuning for Video LLMs | ✓ Link | 74.1 | 52.9 | | | 83.6 | 62.1 | TempSamp-R1 (CoT) | 2025-09-22 |
| SARL-STG: A Spatially Aware Reinforcement Learning Framework for Refining MLLMs in Spatio-Temporal Video Grounding | | 73.6 | 49.1 | | | | 61.3 | SARL-STG | 2026-06-03 |
| The Devil is in the Spurious Correlations: Boosting Moment Retrieval with Dynamic Learning | ✓ Link | 73.49 | 53.01 | | | | | TD-DETR (InternVideo2) (extra data) | 2025-01-13 |
| Invert4TVG: A Temporal Video Grounding Framework with Inversion Tasks Preserving Action Understanding Ability | ✓ Link | 72.5 | 51.4 | | | 83.0 | | Invert4TVG 7B (extra data) | 2025-08-10 |
| TimeZero: Temporal Video Grounding with Reasoning-Guided LVLM | ✓ Link | 72.2 | 50.1 | | | 82.8 | 60.9 | Time-R1-7B (fine-tuned) (extra data) | 2025-03-17 |
| Sparse-Dense Side-Tuner for efficient Video Temporal Grounding | ✓ Link | 72.0 | 52.6 | | | | 61.2 | SDST (InternVideo2) | 2025-07-10 |
| See More, Store Less: Memory-Efficient Resolution for Video Moment Retrieval | | 71.26 | 49.81 | | | | | SMORE | 2026-01-14 |
| Saliency-Guided DETR for Moment Retrieval and Highlight Detection | ✓ Link | 71.10 | 52.80 | | | | | SG-DETR (w/ PT) | 2024-10-02 |
| LLaVA-MR: Large Language-and-Vision Assistant for Video Moment Retrieval | ✓ Link | 70.65 | 49.58 | | | | | LLaVA-MR | 2024-11-21 |
| Incentivizing Versatile Video Reasoning in MLLMs via Data-Efficient Reinforcement Learning | ✓ Link | 70.4 | 45.7 | | | 81.9 | 59.1 | VideoReasoner | 2026-06-03 |
| FlashVTG: Feature Layering and Adaptive Score Handling Network for Video Temporal Grounding | ✓ Link | 70.32 | 49.87 | | | | | FlashVTG | 2024-12-18 |
| Saliency-Guided DETR for Moment Retrieval and Highlight Detection | ✓ Link | 70.20 | 49.50 | | | | | SG-DETR | 2024-10-02 |
| InternVideo2: Scaling Foundation Models for Multimodal Video Understanding | ✓ Link | 70.03 | 48.95 | | | | | InternVideo2-6B | 2024-03-22 |
| VideoLoom: A Video Large Language Model for Joint Spatial-Temporal Understanding | ✓ Link | 70.0 | 48.3 | | | | | VideoLoom-8B | 2026-01-12 |
| DTOS: Dynamic Time Object Sensing with Large Multimodal Model | ✓ Link | 69.81 | 43.2 | | | 81.29 | 58.0 | DTOS-TCS | 2025-06-11 |
| DeCafNet: Delegate and Conquer for Efficient Temporal Grounding in Long Videos | ✓ Link | 68.79 | 47.55 | 91.53 | 72.96 | | | DeCaf-Grounder (extra data) | 2025-05-22 |
| Measure Twice, Cut Once: Grasping Video Structures and Event Semantics with LLMs for Video Temporal Localization | ✓ Link | 68.5 | 41.6 | | | 82.3 | | MeCo (QWen2VL 7B) (extra data) | 2025-03-12 |
| Knowing Where to Focus: Event-aware Transformer for Video Grounding | ✓ Link | 68.47 | 44.92 | | | | | EaTR | 2023-08-14 |
| InternVideo2: Scaling Foundation Models for Multimodal Video Understanding | ✓ Link | 68.36 | 45.03 | | | | | InternVideo2-1B | 2024-03-22 |
| Temporal-Aware Reasoning Optimization for Video Temporal Grounding | ✓ Link | 67.8 | 41.9 | | | 83.2 | | TaRO 8B (extra data) | 2026-06-08 |
| SlotVTG: Object-Centric Adapter for Generalizable Video Temporal Grounding | | 67.6 | 46.7 | | | 79.5 | 58.3 | SlotVTG 7B (extra data) | 2026-03-26 |
| TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning | ✓ Link | 67.1 | 43.0 | | | | | VideoChat-T (FT) | 2024-10-25 |
| VideoTG-R1: Boosting Video Temporal Grounding via Curriculum Reinforcement Learning on Reflected Boundary Annotations | ✓ Link | 64.8 | 39.2 | | | 79.5 | 55.9 | VideoTG-R1 (extra data) | 2025-10-27 |
| UniMD: Towards Unifying Moment Retrieval and Temporal Action Detection | ✓ Link | 63.98 | 44.46 | 91.94 | 67.72 | | | UniMD+Sync. | 2024-04-07 |
| SpaceVLLM: Endowing Multimodal Large Language Model with Spatio-Temporal Video Grounding Capability | ✓ Link | 63.6 | 38.5 | | | | | SpaceVLLM 7B (extra data) | 2025-03-18 |
| OmniVTG: A Large-Scale Dataset and Training Paradigm for Open-World Video Temporal Grounding | ✓ Link | 63.2 | 37.0 | | | 78.3 | | OmniVTG-7B (zero-shot) (extra data) | 2026-04-28 |
| CVA: Context-aware Video-text Alignment for Video Temporal Grounding | | 62.61 | 40.78 | | | 74.19 | 53.35 | CVA (extra data) | 2026-03-26 |
| LD-DETR: Loop Decoder DEtection TRansformer for Video Moment Retrieval and Highlight Detection | ✓ Link | 62.58 | 41.56 | | | 73.92 | 53.44 | LD-DETR | 2025-01-18 |
| VideoLights: Feature Refinement and Cross-Task Alignment Transformer for Joint Video Highlight Detection and Moment Retrieval | ✓ Link | 61.96 | 41.05 | | | 73.33 | 52.94 | VideoLights-B-pt | 2024-12-02 |
| GIRL-DETR: Gradient-Isolated Reinforcement Learning for Video Moment Retrieval | ✓ Link | 61.75 | 37.47 | | | | | GIRL-DETR | 2026-05-30 |
| TRACE: Temporal Grounding Video LLM via Causal Event Modeling | ✓ Link | 61.7 | 41.4 | | | | | TRACE (extra data) | 2024-10-08 |
| Keeping the Evidence Chain: Semantic Evidence Allocation for Training-Free Token Pruning in Video Temporal Grounding | ✓ Link | 61.08 | 33.17 | | | 76.91 | 52.31 | SemVID (Qwen3-VL-4B, 25% visual-token retention) | 2026-03-05 |
| HERO: Hierarchical Embedding-Refinement for Open-Vocabulary Temporal Sentence Grounding in Videos | | 61.05 | 41.29 | | | | | HERO (extra data) | 2026-03-06 |
| UnLoc: A Unified Framework for Video Localization Tasks | ✓ Link | 60.8 | 38.4 | 88.2 | 61.1 | | | UnLoc-L | 2023-08-21 |
| T2SGrid: Temporal-to-Spatial Gridification for Video Temporal Grounding | | 60.6 | 35.9 | | | 76.9 | 53.2 | Qwen2-VL-7B + T2SGrid-FT (extra data) | 2026-03-07 |
| Beyond Uncertainty: Evidential Deep Learning for Robust Video Temporal Grounding | ✓ Link | 60.2 | 38.0 | | | 71.6 | 51.6 | SRAM-Base | 2024-08-29 |
| SE-DETR: Explicit Semantic Exploration for Generalizability and Distinguishability in Video Temporal Grounding | ✓ Link | 60.2 | 39.7 | | | 71.1 | 51.2 | SE-DETR+ | 2026-09-08 |
| UniVTG: Towards Unified Video-Language Temporal Grounding | ✓ Link | 60.19 | 38.55 | | | 72.63 | 52.17 | UniVTG w/ PT (extra data) | 2023-07-31 |
| BAM-DETR: Boundary-Aligned Moment Detection Transformer for Temporal Sentence Grounding in Videos | ✓ Link | 59.95 | 39.38 | | | | | BAM-DETR | 2023-11-30 |
| $R^2$-Tuning: Efficient Image-to-Video Transfer Learning for Video Temporal Grounding | ✓ Link | 59.78 | 37.02 | | | 70.91 | 50.86 | R2-Tuning | 2024-03-31 |
| Background-aware Moment Detection for Video Moment Retrieval | ✓ Link | 59.48 | 38.33 | | | | | BM-DETR | 2023-06-05 |
| Bridging the Gap: A Unified Video Comprehension Framework for Moment Retrieval and Highlight Detection | ✓ Link | 59.25 | 36.64 | | | | | UVCOM | 2023-11-28 |
| VideoMind: A Chain-of-LoRA Agent for Long Video Reasoning | ✓ Link | 59.1 | 31.2 | | | 73.5 | 50.2 | VideoMind 7B (zero-shot) (extra data) | 2025-03-17 |
| Moment Quantization for Video Temporal Grounding | ✓ Link | 58.84 | 38.84 | | | 70.97 | 51.10 | MQVTG (extra data) | 2025-04-03 |
| DSE-VTG: Dual-Side Enhancement for Training-Free Video Temporal Grounding | | 58.58 | 32.23 | | | 75.65 | 51.30 | DSE-VTG | 2026-09-08 |
| Correlation-Guided Query-Dependency Calibration for Video Temporal Grounding | ✓ Link | 58.44 | 36.34 | | | | | CG-DETR | 2023-11-15 |
| Locate Anything in Videos: Rethinking Efficient Generative Spatio-Temporal Video Grounding | ✓ Link | 58.4 | 30.1 | | | 78.9 | 52.4 | PTD Qwen3-VL-4B | 2026-08-28 |
| Prior Knowledge Integration via LLM Encoding and Pseudo Event Regulation for Video Moment Retrieval | ✓ Link | 58.31 | 36.49 | | | | | LLMEPET | 2024-07-21 |
| HawkEye: Training Video-Text LLMs for Grounding Text in Videos | ✓ Link | 58.3 | 28.8 | | | 72.5 | 49.3 | HawkEye 7B (fine-tuned) (extra data) | 2024-03-15 |
| MomentSeg: Moment-Centric Sampling for Enhanced Referring Video Object Segmentation | | 58.2 | 25.8 | | | 76.1 | 50.2 | MomentSeg-3B | 2026-09-08 |
| UnLoc: A Unified Framework for Video Localization Tasks | ✓ Link | 58.1 | 35.4 | 87.4 | 59.1 | | | UnLoc-B | 2023-08-21 |
| VTG-LLM: Integrating Timestamp Knowledge into Video LLMs for Enhanced Video Temporal Grounding | ✓ Link | 57.8 | 33.9 | | | | | VTG-LLM (τ=0.1) (extra data) | 2024-05-22 |
| GroundVTS: Visual Token Sampling in Multimodal Large Language Models for Video Temporal Grounding | ✓ Link | 57.5 | 34.2 | | | 71.5 | 50.1 | GroundVTS-Q (extra data) | 2026-04-02 |
| Query-Dependent Video Representation for Moment Retrieval and Highlight Detection | ✓ Link | 57.31 | 32.55 | | | | | QD-DETR (Only Video) | 2023-03-24 |
| Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding | ✓ Link | 57.18 | 36.05 | | | | | video-mamba-suite | 2024-03-14 |
| Multi-Modal Interaction Graph Convolutional Network for Temporal Language Localization in Videos | ✓ Link | 57.10 | 34.54 | | | | | MIGCN (I3D) | 2021-10-12 |
| Task-Driven Exploration: Decoupling and Inter-Task Feedback for Joint Moment Retrieval and Highlight Detection | ✓ Link | 56.51 | 33.66 | | | | | TaskWeave (VGG) | 2024-04-14 |
| Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders | ✓ Link | 56.0 | | | | | | Penguin-VL 8B | 2026-03-06 |
| Grounded-VideoLLM: Sharpening Fine-grained Temporal Grounding in Video Large Language Models | ✓ Link | 55.9 | 33.2 | | | 70.2 | 49.4 | Grounded-VideoLLM Phi-3.5 (task-tuned) (extra data) | 2024-10-04 |
| MomentSeg: Moment-Centric Sampling for Enhanced Video Pixel Understanding | ✓ Link | 55.9 | 25.3 | | | 78.9 | 50.7 | MomentSeg-3B (θ=0.3) (extra data) | 2025-10-10 |
| QVHighlights: Detecting Moments and Highlights in Videos via Natural Language Queries | ✓ Link | 55.65 | 34.17 | | | | | Moment-DETR w/ PT (on 10K HowTo100M videos) | 2021-07-20 |
| Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders | ✓ Link | 54.5 | | | | | | Penguin-VL 2B | 2026-03-06 |
| QVHighlights: Detecting Moments and Highlights in Videos via Natural Language Queries | ✓ Link | 53.63 | 31.37 | | | | | Moment-DETR | 2021-07-20 |
| Training-free Uncertainty Guidance for Complex Visual Tasks with MLLMs | ✓ Link | 52.5 | 32.6 | | | 74.7 | 51.0 | UG-Ground (InternVideo2.5-8B) | 2025-10-01 |
| DART: Difficulty-Adaptive Routing for Zero-Shot Video Temporal Grounding | | 52.04 | 29.45 | | | 70.98 | 48.93 | DART | 2026-07-01 |
| Foresee-to-Ground: From Predictive Temporal Perception to Evidence-Driven Reasoning for Video Temporal Grounding | ✓ Link | 50.4 | 25.7 | | | 71.3 | 47.2 | F2G (Qwen3-VL-8B, zero-shot) (extra data) | 2026-05-21 |
| UMT: Unified Multi-modal Transformers for Joint Video Moment Retrieval and Highlight Detection | ✓ Link | 49.35 | 26.16 | 89.41 | 54.95 | | | UMT (VO) | 2022-03-23 |
| Explore-And-Match: Bridging Proposal-Based and Proposal-Free With Transformer for Sentence Grounding in Videos | ✓ Link | 49.11 | 26.59 | 88.5 | 55.99 | | 47.13 | LVTR-CLIP (extra data) | 2022-01-25 |
| TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning | ✓ Link | 48.7 | 24.0 | | | | 45.43 | VideoChat-T (ZS) | 2024-10-25 |
| UMT: Unified Multi-modal Transformers for Joint Video Moment Retrieval and Highlight Detection | ✓ Link | 48.31 | 29.25 | 88.79 | 56.08 | | | UMT (VA) | 2022-03-23 |
| SnAG: Scalable and Accurate Video Grounding | ✓ Link | 46.26 | 33.52 | 92.55 | 71.94 | | | SnAG | 2024-04-02 |
| LLaVA-ST: A Multimodal Large Language Model for Fine-Grained Spatial-Temporal Understanding | ✓ Link | 44.8 | 23.4 | | | 63.1 | 42.4 | LLaVA-ST (extra data) | 2025-01-14 |
| SimVTP: Simple Video Text Pre-training with Masked Autoencoders | | 44.7 | 26.3 | 83.7 | 55.1 | | | SimVTP | 2022-12-07 |
| ARC-Hunyuan-Video-7B: Structured Video Comprehension of Real-World Shorts | ✓ Link | | | | | | 54.8 | ARC-Hunyuan-Video-7B (extra data) | 2025-07-28 |
| Text-Visual Prompting for Efficient 2D Temporal Video Grounding | ✓ Link | 44.39 | 21.51 | | | 65.92 | | TVP (extra data) | 2023-03-09 |
| Text-Visual Prompting for Efficient 2D Temporal Video Grounding | ✓ Link | 44.39 | | | | | | TVP | 2023-03-09 |
| VTG-GPT: Tuning-Free Zero-Shot Video Temporal Grounding with GPT | ✓ Link | 43.68 | 25.94 | | | 59.48 | 39.81 | VTG-GPT (zero-shot) (extra data) | 2024-03-04 |
| VideoLLM Knows When to Speak: Enhancing Time-Sensitive Video Comprehension with Video-Text Duet Interaction Format | ✓ Link | 42.4 | 18.0 | | | | | MMDuet (zero-shot) (extra data) | 2024-11-27 |
| Number it: Temporal Grounding Videos like Flipping Manga | ✓ Link | 42.0 | 20.6 | | | 63.8 | 41.4 | LongVA-7B-DPO + NumPro-FT (extra data) | 2024-11-15 |
| D3G: Exploring Gaussian Prior for Temporal Sentence Grounding with Glance Annotation | ✓ Link | 41.64 | 19.60 | 79.25 | 49.30 | | | D3G (glance-supervised) | 2023-08-08 |
| Learning 2D Temporal Adjacent Networks for Moment Localization with Natural Language | ✓ Link | 39.81 | 23.25 | 79.33 | 52.15 | | | 2D-TAN (Conv) | 2019-12-08 |
| Mitigating Modality and Language-Style Gaps for Zero-Shot Video Moment Retrieval | | 39.7 | 21.0 | | | 62.7 | 41.9 | Self-SiMS | 2026-07-21 |
| Seq2Time: Sequential Knowledge Transfer for Video LLM Temporal Grounding | | 31.2 | 13.7 | | | | | TimeChat + Seq2Time (extra data) | 2024-11-25 |
| Joint Moment Retrieval and Highlight Detection Via Natural Language Queries | ✓ Link | 25.97 | 13.25 | 84.87 | 40.38 | | | VisionaryVid (video + audio) | 2023-05-08 |
| VideoTemp-o3: Harmonizing Temporal Grounding and Video Understanding in Agentic Thinking-with-Videos | | | 33.0 | | | | 57.8 | VideoTemp-o3-7B-RL (extra data) | 2026-02-08 |
| Reinforcement Learning Tuning for VideoLLMs: Reward Design and Data Efficiency | ✓ Link | | | | | | 58.0 | Temporal-RLT (4096 video tokens) (extra data) | 2025-06-02 |
| TimeChat-Captioner: Scripting Multi-Scene Videos with Time-Aware and Structural Audio-Visual Captions | ✓ Link | | | | | | 58.8 | TimeChat-Captioner (7B, GRPO) | 2026-02-09 |
| VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering Twice | ✓ Link | | | | | | 63.7 | VideoAuto-R1 (Qwen3-VL-8B) | 2026-01-08 |
| VisionCoach: Reinforcing Grounded Video Reasoning via Visual-Perception Prompting | ✓ Link | | | | | | 42.7 | VisionCoach-7B | 2026-03-15 |
| ResAdapt: Adaptive Resolution for Efficient Multimodal Reasoning | ✓ Link | | | | | | 42.0 | ResAdapt (Qwen2.5-VL-7B, 128 frames) | 2026-03-30 |
| ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning | ✓ Link | | | | | | 50.1 | ParaVT (Qwen3-VL-8B) | 2026-05-19 |
| DRIFT: A Residual Flow Adapter for Decoding Continuous Outputs in Vision-Language Models | ✓ Link | | | | | | 43.8 | DRIFT (ET-Chat) | 2026-06-04 |
| Less from More: Reinforcing Sparse Video Reasoning from Dense References | ✓ Link | | | | | | 54.8 | SAVER-4B (extra data) | 2026-10-07 |
| EventLens: Event-Structure Reinforcement Learning for Video Understanding | ✓ Link | | | | | | 48.25 | EventLens-8B | 2026-10-12 |