Video-Text Retrieval on MSR-VTT
78.6R@1MAVIS
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MAVISMethod Category=Agentic Methods2026.06 | 78.6 | — | — | — | 94.9 | 96.3 | — | |
| MERLIN (Round 5)Method Category=Agentic Methods2026.06 | 78 | — | — | — | 94.2 | 96.8 | — | |
| MERLIN (Round 3)Method Category=Agentic Methods2026.06 | 72.6 | — | — | — | 91.8 | 95.6 | — | |
| InternVideo2-6BMethod Category=Large-scale Video Foundation Models2026.06 | 55.9 | — | — | — | 78.3 | 85.1 | — | |
| T-MASSMethod Category=Supervised / Fine-tuned2026.06 | 52.7 | — | — | — | 77.1 | 85.6 | — | |
| Cap4VideoMethod Category=Supervised / Fine-tuned2026.06 | 51.4 | — | — | — | 75.7 | 83.9 | — | |
| Marengo-2.6Method Category=Large-scale Video Foundation Models2026.06 | 49.35 | — | — | — | 73.47 | — | — | |
| X-CLIPMethod Category=Supervised / Fine-tuned2026.06 | 49.3 | — | — | — | 75.8 | 84.8 | — | |
| VASTMethod Category=Large-scale Video Foundation Models2026.06 | 49.3 | — | — | — | 68.3 | 73.9 | — | |
| TeachCLIPMethod Category=Supervised / Fine-tuned2026.06 | 46.8 | — | — | — | 74.3 | 82.6 | — | |
| LanguageBind-HMethod Category=Large-scale Video Foundation Models2026.06 | 44.8 | — | — | — | 70 | 78.7 | — | |
| CLIP4ClipMethod Category=Supervised / Fine-tuned2026.06 | 44.5 | — | — | — | 71.4 | 81.6 | — | |
| MERLIN (Round 0)Method Category=Agentic Methods2026.06 | 44.4 | — | — | — | 67.6 | 76.2 | — | |
| VideoPrism-gMethod Category=Large-scale Video Foundation Models2026.06 | 39.7 | — | — | — | 63.7 | — | — | |
| InternVideoZero-shot=true, WiSE-FT=true2023.10 | 36.2 | — | — | — | 60.3 | 69.7 | 3 | |
| OmniVLZero-shot=true, Reranking (Matching loss)=true2023.10 | 34.6 | — | — | — | 58.4 | 66.6 | — | |
| BridgeFormerZero-shot=true2023.10 | 33.2 | — | — | — | 58 | 68.6 | 4 | |
| InternVideo (ViT Backbone)Zero-shot=true, Backbone=ViT, WiSE-FT=true2023.10 | 32.6 | — | — | — | 54.9 | 65.5 | 4 | |
| COXSetting=Semi-Supervised, IM Perceptor=LanguageBind2025.12 | 28.7 | — | — | — | 49.5 | — | — | |
| COXSetting=Semi-Supervised, IM Perceptor=ImageBind2025.12 | 26.5 | — | — | — | 48.8 | — | — | |
| VIOLETZero-shot=true2023.10 | 25.9 | — | — | — | 49.5 | 59.7 | — | |
| FrozenZero-shot=true2023.10 | 24.7 | — | — | — | 46.9 | 57.2 | 7 | |
| GRU+SSA2021.05 | 24.4 | — | — | — | 49.3 | — | — | |
| ALPROZero-shot=true2023.10 | 24.1 | — | — | — | 44.7 | 55.4 | 8 | |
| COXSetting=Unsupervised, IM Perceptor=LanguageBind2025.12 | 15.4 | — | — | — | 35 | — | — | |
| COXSetting=Unsupervised, IM Perceptor=ImageBind2025.12 | 15.2 | — | — | — | 34.8 | — | — | |
| GRU+ResNeXt-101Backbone=ResNeXt-1012021.05 | 12.1 | — | — | — | 35 | — | — | |
| VideoCLIPZero-shot=true2023.10 | 10.4 | — | — | — | 22.2 | 30 | — | |
| JSFusion2021.05 | 10.2 | — | — | — | 31.2 | — | — | |
| Kaufman et al.2021.05 | 4.7 | — | — | — | 16.6 | — | — | |
| CT-SAN2021.05 | 4.4 | — | — | — | 16.6 | — | — | |
| C+LSTM+SA2021.05 | 4.2 | — | — | — | 12.9 | — | — | |
| VSE-LSTM2021.05 | 3.8 | — | — | — | 12.7 | — | — | |
| Random2021.05 | 0.1 | — | — | — | 0.5 | — | — | |
| AdaLoRABackbone=ViT+Text Transf., Params. (M)=1.2, Mem. (G)=42.0, QPS=40.062026.04 | — | 39.2 | 39.6 | 368.5 | — | — | — | |
| AdapterBackbone=ViT+Text Transf., Params. (M)=5.2, Mem. (G)=41.2, QPS=39.172026.04 | — | 38.3 | 39.6 | 364.3 | — | — | — | |
| BitFitBackbone=ViT+Text Transf., Params. (M)=0.1, Mem. (G)=42.0, QPS=43.122026.04 | — | 38.1 | 40.6 | 370.8 | — | — | — | |
| FacTBackbone=ViT+Text Transf., Params. (M)=0.8, Mem. (G)=40.82026.04 | — | 38.7 | 39.8 | 367.2 | — | — | — | |
| Fully-FTParams (M)=151.3, Training Memory (G)=12.2 * 4, Test Memory (G)=1.12, Architecture=CLIP4Clip2023.08 | — | 42.8 | 42.1 | 389.2 | — | — | — | |
| Fully-FTBackbone=ViT+Text Transf., Params. (M)=151.3, Mem. (G)=48.8, QPS=44.072026.04 | — | 42.8 | 42.1 | 389.2 | — | — | — | |
| Fully-FTParams (M)=151.3, Mem (G)=48.82026.04 | — | 42.8 | 42.1 | 389.2 | — | — | — | |
| Fully-FTParams. (M)=151.3, Mem. (G)=48.82026.04 | — | 42.8 | 42.1 | 389.2 | — | — | — | |
| LoRABackbone=ViT+Text Transf., Params. (M)=1.3, Mem. (G)=40.8, QPS=41.552026.04 | — | 38.8 | 39.9 | 366.8 | — | — | — | |
| LSTParams (M)=11.2, Training Memory (G)=8.0 * 4, Test Memory (G)=1.15, Architecture=CLIP4Clip2023.08 | — | 37 | 37.8 | 356.7 | — | — | — | |
| LSTBackbone=ViT+Text Transf., Params. (M)=11.2, Mem. (G)=32.0, QPS=19.852026.04 | — | 37 | 37.8 | 356.7 | — | — | — | |
| LSTParams (M)=11.2, Mem (G)=32.02026.04 | — | 37 | 37.8 | 356.7 | — | — | — | |
| LSTParams. (M)=11.2, Mem. (G)=32.02026.04 | — | 37 | 37.8 | 356.7 | — | — | — | |
| MDPDBackbone=ViT+Text Transf., Params. (M)=9.7, Mem. (G)=14.0, QPS=43.522026.04 | — | 40.2 | 40.7 | 365.8 | — | — | — | |
| Ours†Params. (M)=10.1, Mem. (G)=14.52026.04 | — | 39.9 | 41.1 | 365.5 | — | — | — | |
| Ours‡Params. (M)=10.1, Mem. (G)=14.62026.04 | — | 40.3 | 41.3 | 366.9 | — | — | — | |
| Ours♠Params (M)=9.7, Mem (G)=14.2, QPS=43.522026.04 | — | 40.6 | 41.2 | 367.2 | — | — | — | |
| Ours♡Params (M)=9.7, Mem (G)=14.0, QPS=43.652026.04 | — | 40.2 | 40.7 | 365.8 | — | — | — | |
| PartiallyBackbone=ViT+Text Transf., Params. (M)=0.7, Mem. (G)=7.6, QPS=44.482026.04 | — | 36.4 | 37 | 353.9 | — | — | — | |
| Partially⇓Params (M)=0.7, Training Memory (G)=1.9 * 4, Test Memory (G)=1.12, Architecture=CLIP4Clip2023.08 | — | 36.4 | 37 | 353.9 | — | — | — | |
| PromptBackbone=ViT+Text Transf., Params. (M)=0.2, Mem. (G)=42.8, QPS=33.742026.04 | — | 36.8 | 37.5 | 358.8 | — | — | — | |
| SHERLParams (M)=9.6, Mem (G)=13.6, QPS=32.532026.04 | — | 39.2 | 40.6 | 363.7 | — | — | — | |
| SHERLParams. (M)=9.6, Mem. (G)=13.62026.04 | — | 39.2 | 40.6 | 363.7 | — | — | — | |
| SSFBackbone=ViT+Text Transf., Params. (M)=0.5, Mem. (G)=39.2, QPS=37.912026.04 | — | 40.2 | 41.8 | 376.6 | — | — | — | |
| UniPTParams (M)=9.6, Training Memory (G)=3.4 * 4, Test Memory (G)=1.13, Architecture=CLIP4Clip2023.08 | — | 38.9 | 39.3 | 361.3 | — | — | — | |
| UniPTBackbone=ViT+Text Transf., Params. (M)=9.6, Mem. (G)=13.6, QPS=27.032026.04 | — | 38.9 | 39.3 | 361.3 | — | — | — | |
| UniPTParams (M)=9.6, Mem (G)=13.6, QPS=27.032026.04 | — | 38.9 | 39.3 | 361.3 | — | — | — | |
| UniPTParams. (M)=9.6, Mem. (G)=13.62026.04 | — | 38.9 | 39.3 | 361.3 | — | — | — |