Text-driven Motion Generation on HumanML3D (test)
58.1R-Precision@1SALAD
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SALADFinetuned on DeCompML=false2025.12 | 58.1 | 76.9 | 85.7 | 0.076 | 2.649 | — | — | 9.696 | — | |
| SALADType=Text-to-Motion Only2025.12 | 58.1 | 76.9 | 85.7 | 0.076 | 2.649 | 1.751 | — | 9.696 | — | |
| SALADFinetuned on DeCompML=true2025.12 | 58 | 76.9 | 85.7 | 0.06 | 2.651 | — | — | 9.379 | — | |
| MotionLCMsteps=1-step, Retrieval/Refinement=WINRO2026.07 | 58 | 77.5 | 86.1 | 0.072 | 2.609 | — | — | 9.242 | 6.87 | |
| MotionLCMsteps=1-step, Retrieval/Refinement=WIN-RLoRA2026.07 | 58 | 77.2 | 86 | 0.083 | 2.649 | — | — | 9.55 | 0.043 | |
| MotionLCMsteps=1-step, Retrieval/Refinement=WIN-FAR2026.07 | 57.2 | 76.7 | 85.5 | 0.083 | 2.674 | — | — | 9.474 | 0.033 | |
| MotionLCMsteps=1-step, Retrieval/Refinement=WIN-R2026.07 | 56.6 | 76.2 | 85.2 | 0.091 | 2.699 | — | — | 9.424 | — | |
| OmniMoGen-ThinkType=Multi-Task2025.12 | 55 | 74.2 | 83.4 | 0.061 | 2.729 | 1.673 | — | 9.73 | — | |
| MotionLCMsteps=1-step, Retrieval/Refinement=None2026.07 | 54.8 | 74.1 | 83.5 | 0.093 | 2.763 | — | — | 9.566 | 0.029 | |
| MDM-50stepRetrieval/Refinement=WIN-RLoRA2026.07 | 54.6 | 74.1 | 83.2 | 0.102 | 2.846 | — | — | 9.577 | 0.28 | |
| MotionGPT3Type=Multi-Task2025.12 | 54.3 | 73.5 | 82.8 | 0.217 | 2.793 | 1.366 | — | 9.662 | — | |
| LG-Tok#Tokens=2362026.02 | 54.2 | 73.6 | 83 | 0.057 | 2.997 | 1.54 | 0.669 | — | — | |
| MDM-50stepRetrieval/Refinement=WINRO2026.07 | 53.9 | 72.7 | 81.7 | 0.115 | 2.883 | — | — | 9.657 | 81.408 | |
| LG-Tok-mid#Tokens=1602026.02 | 53.7 | 72.9 | 82.1 | 0.109 | 3.061 | 1.674 | 0.664 | — | — | |
| EnergyMoGenFinetuned on DeCompML=true2025.12 | 52.6 | 71.8 | 81.8 | 0.147 | 2.884 | — | — | 9.392 | — | |
| OmniMoGenType=Multi-Task2025.12 | 52.5 | 71.9 | 81.4 | 0.093 | 2.761 | 1.464 | — | 9.692 | — | |
| EnergyMoGenFinetuned on DeCompML=false2025.12 | 52.3 | 71.5 | 81.5 | 0.188 | 2.915 | — | — | 9.488 | — | |
| LG-Tok-mini#Tokens=1042026.02 | 52.1 | 71.5 | 81.1 | 0.085 | 3.113 | 1.728 | 0.655 | — | — | |
| MoMaskType=Text-to-Motion Only2025.12 | 52.1 | 71.3 | 80.7 | 0.045 | 2.958 | 1.241 | — | 9.62 | — | |
| MoMaskcitation=[26]2024.10 | 52.1 | 71.3 | 80.7 | 0.045 | 2.958 | 1.241 | — | 9.62 | — | |
| MoSaVenue=Arxiv' 25, #Tokens=236, reimplementation=meng67 representation2026.02 | 51.8 | 71.2 | 80.9 | 0.064 | 3.15 | 1.789 | 0.657 | — | — | |
| MG-MotionLLMType=Multi-Task2025.12 | 51.6 | 70.6 | 80.2 | 0.303 | 2.952 | 2.125 | — | 9.96 | — | |
| Motion-R1Type=Text-to-Motion Only2025.12 | 51.5 | 71.9 | 81.8 | 0.201 | 2.854 | 2.317 | — | 10.026 | — | |
| MotionLLMType=Multi-Task2025.12 | 51.5 | — | 80.1 | 0.23 | 2.967 | — | — | 9.908 | — | |
| LaMoGenComposer=Human2026.03 | 51.3 | 70.4 | 81.3 | 0.206 | 2.993 | 0.973 | — | 9.635 | — | |
| MDM-50stepRetrieval/Refinement=WIN-FAR2026.07 | 51.2 | 70.6 | 80.4 | 0.248 | 3.013 | — | — | 10.295 | 0.268 | |
| RealType=Real Data2025.12 | 51.1 | 70.3 | 79.7 | 0.002 | 2.974 | — | — | 9.503 | — | |
| Real data2026.03 | 51.1 | 70.3 | 79.7 | 0.002 | 2.974 | — | — | 9.503 | — | |
| Real2024.10 | 51.1 | 70.3 | 79.7 | 0.002 | 2.974 | — | — | 9.503 | — | |
| GTDescription=Ground Truth2026.07 | 51.1 | 70.4 | 79.8 | 0.002 | 2.97 | — | — | 9.458 | — | |
| StableMoFusionVenue=ACM MM' 24, reimplementation=meng67 representation2026.02 | 51 | 71 | 81 | 0.177 | 3.182 | 1.969 | 0.654 | — | — | |
| ReMoDiff2026.03 | 51 | 69.8 | 79.5 | 0.103 | 2.974 | 1.795 | — | 9.018 | — | |
| ReMoDiffusecitation=[29]2024.10 | 51 | 69.8 | 79.5 | 0.103 | 2.974 | 1.795 | — | 9.018 | — | |
| ParTYPart=Arms2026.03 | 50.6 | — | 80.2 | 0.133 | 3.079 | — | — | — | — | |
| MDM-50stepRetrieval/Refinement=WIN-R2026.07 | 50.5 | 70.2 | 79.9 | 0.25 | 3.047 | — | — | 10.303 | — | |
| MotionChainType=Multi-Task2025.12 | 50.4 | 69.5 | 79 | 0.248 | 3.033 | 1.715 | — | 9.47 | — | |
| GraphMotioncitation=[71]2024.10 | 50.4 | 69.9 | 78.5 | 0.116 | 3.07 | 2.766 | — | 9.692 | — | |
| CoMo2026.03 | 50.2 | 69.2 | 79 | 0.262 | 3.032 | 1.013 | — | 9.936 | — | |
| Real motions2026.02 | 50.1 | 69.6 | 79.2 | 0 | 3.251 | — | 0.639 | — | — | |
| MARDMVenue=CVPR' 252026.02 | 50 | 69.5 | 79.5 | 0.114 | 3.27 | 2.231 | 0.642 | — | — | |
| AttT2Mcitation=[70]2024.10 | 49.9 | 69 | 78.6 | 0.112 | 3.038 | 2.452 | — | 9.7 | — | |
| T2M-GPTFinetuned on DeCompML=true2025.12 | 49.7 | 68.5 | 78 | 0.102 | 3.017 | — | — | 9.748 | — | |
| MotionGPT-2Type=Multi-Task2025.12 | 49.6 | 69.1 | 78.2 | 0.191 | 3.08 | 2.137 | — | 9.86 | — | |
| KP2026.03 | 49.6 | — | — | 0.275 | — | 2.218 | — | 9.975 | — | |
| MotionGPT-22024.10 | 49.6 | 69.1 | 78.2 | 0.191 | 3.08 | 2.137 | — | 9.86 | — | |
| MotionGPTType=Multi-Task, citation=[12]2025.12 | 49.2 | 68.1 | 77.8 | 0.232 | 3.096 | 2.008 | — | 9.528 | — | |
| MotionGPT2026.03 | 49.2 | 68.1 | 77.8 | 0.232 | 3.096 | 2.008 | — | 9.528 | — | |
| MotionGPTcitation=[27]2024.10 | 49.2 | 68.1 | 73.3 | 0.232 | 3.096 | 2.008 | — | 9.528 | — | |
| T2M-GPTFinetuned on DeCompML=false2025.12 | 49.1 | 68 | 77.5 | 0.116 | 3.118 | — | — | 9.761 | — | |
| MotionDiffuseType=Text-to-Motion Only2025.12 | 49.1 | 68.1 | 78.2 | 0.63 | 3.113 | 1.553 | — | 9.41 | — | |
| T2M-GPTType=Text-to-Motion Only2025.12 | 49.1 | 68 | 77.5 | 0.116 | 3.118 | 1.856 | — | 9.761 | — | |
| MoDiff2026.03 | 49.1 | 68.1 | 78.2 | 0.63 | 3.113 | 1.533 | — | 9.41 | — | |
| LaMoGenComposer=GPT4.12026.03 | 49.1 | 69.4 | 79.6 | 0.252 | 3.087 | 1.131 | — | 9.124 | — | |
| MotionDiffusecitation=[17]2024.10 | 49.1 | 68.1 | 78.2 | 0.63 | 3.113 | 1.553 | — | 9.41 | — | |
| T2M-GPTcitation=[24]2024.10 | 49.1 | 68 | 77.5 | 0.116 | 3.118 | 1.856 | — | 9.761 | — | |
| MoMaskVenue=CVPR' 24, #Tokens=2942026.02 | 49 | 68.7 | 78.6 | 0.116 | 3.353 | 1.263 | 0.637 | — | — | |
| MMMVenue=CVPR' 24, #Tokens=492026.02 | 48.7 | 68.3 | 78.2 | 0.132 | 3.359 | 1.241 | 0.635 | — | — | |
| MotionLLMcitation=[65]2024.10 | 48.2 | 67.2 | 77 | 0.491 | 3.138 | — | — | 9.838 | — | |
| MLDType=Text-to-Motion Only2025.12 | 48.1 | 67.3 | 77.2 | 0.473 | 3.196 | 2.413 | — | 9.724 | — | |
| MLDcitation=[25]2024.10 | 48.1 | 67.3 | 77.2 | 0.473 | 3.169 | 2.413 | — | 9.724 | — | |
| T2M-GPTVenue=CVPR' 23, #Tokens=492026.02 | 47 | 65.9 | 75.8 | 0.335 | 3.505 | 2.018 | 0.607 | — | — | |
| MDM-50stepRetrieval/Refinement=None2026.07 | 46.9 | 66.1 | 76.3 | 0.438 | 3.258 | — | — | 10.025 | 0.265 | |
| ParCoPart=Arms2026.03 | 46.8 | — | 76.7 | 0.215 | 3.326 | — | — | — | — | |
| ParTYPart=Legs2026.03 | 46.3 | — | 75.5 | 0.078 | 3.122 | — | — | — | — | |
| T2MType=Text-to-Motion Only2025.12 | 45.7 | 63.9 | 74 | 1.067 | 3.34 | 2.09 | — | 9.188 | — | |
| Guo et al.2026.03 | 45.7 | 63.9 | 74 | 1.067 | 3.34 | 2.09 | — | 9.188 | — | |
| T2Mcitation=[30]2024.10 | 45.7 | 63.9 | 74 | 1.067 | 3.34 | 2.09 | — | 9.188 | — | |
| LaMoGenComposer=GPT4.1mini2026.03 | 45.3 | 67.9 | 77.9 | 0.561 | 3.717 | 1.263 | — | 8.434 | — | |
| MoMaskPart=Arms2026.03 | 45.2 | — | 76.1 | 0.175 | 3.44 | — | — | — | — | |
| MotionDiffuseVenue=TPAMI' 242026.02 | 45 | 64.1 | 75.3 | 0.778 | 3.49 | 3.179 | 0.606 | — | — | |
| LaMoGenComposer=None2026.03 | 43.8 | 59.1 | 75.5 | 1.091 | 3.999 | 1.421 | — | 8.555 | — | |
| TM2TType=Text-to-Motion Only2025.12 | 42.4 | 61.8 | 72.9 | 1.501 | 3.467 | 2.424 | — | 8.589 | — | |
| TM2Tcitation=[63]2024.10 | 42.4 | 61.8 | 72.9 | 1.501 | 3.467 | 2.424 | — | 8.589 | — | |
| ParCoPart=Legs2026.03 | 40.7 | — | 69.9 | 0.118 | 3.482 | — | — | — | — | |
| MoMaskPart=Legs2026.03 | 40.3 | — | 68.7 | 0.104 | 3.513 | — | — | — | — | |
| MotionGPTType=Multi-Task, citation=[44]2025.12 | 36.4 | 53.3 | 62.9 | 0.805 | 3.914 | 2.473 | — | 9.972 | — | |
| MotionGPTcitation=[1]2024.10 | 36.4 | 53.3 | 62.9 | 0.805 | 3.914 | 2.473 | — | 9.972 | — | |
| MDMType=Text-to-Motion Only2025.12 | 32 | 49.8 | 61.1 | 0.544 | 5.566 | 2.799 | — | 9.559 | — | |
| MDM2026.03 | 32 | 49.8 | 61.1 | 0.544 | 5.566 | 2.799 | — | 9.559 | — | |
| MDMcitation=[69]2024.10 | 31.9 | 49.8 | 61.1 | 0.544 | 5.566 | 2.799 | — | 9.559 | — |