BenchmarkPAW-GEN-10
Environmente02Datum-corner alignment
- Episodes
- DiT-Flow 30 Claude Fable 5.2 (D10) 30 Claude Opus 5.5 (D10) 30 Gemini 4 Argon (D10) 30 GPT-6 Astra (D10) 30 GR00T N1.7 30 MolmoAct 2 30 Pi 0.5 30 Rho 30
- Total duration
- DiT-Flow 8:24 Claude Fable 5.2 (D10) 23:19 Claude Opus 5.5 (D10) 100:31 Gemini 4 Argon (D10) 12:01 GPT-6 Astra (D10) 50:18 GR00T N1.7 8:49 MolmoAct 2 6:43 Pi 0.5 7:06 Rho 9:33
- Mean duration
- DiT-Flow 16.8 s Claude Fable 5.2 (D10) 46.6 s Claude Opus 5.5 (D10) 201.0 s Gemini 4 Argon (D10) 24.0 s GPT-6 Astra (D10) 100.6 s GR00T N1.7 17.6 s MolmoAct 2 13.4 s Pi 0.5 14.2 s Rho 19.1 s
- Success
- DiT-Flow 33% Claude Fable 5.2 (D10) 80% Claude Opus 5.5 (D10) 43% Gemini 4 Argon (D10) 47% GPT-6 Astra (D10) 93% GR00T N1.7 0% MolmoAct 2 43% Pi 0.5 40% Rho 67%
- Manipulation density
- DiT-Flow Share of time manipulating, five definitionsJoints moving98.9%End-effector moving88.2%Gripper engaged0.0%Contact force > 5 N71.6%Gripper engaged or in contact71.6% Claude Fable 5.2 (D10) Share of time manipulating, five definitionsJoints moving89.2%End-effector moving79.3%Gripper engaged42.1%Contact force > 5 N87.6%Gripper engaged or in contact93.0% Claude Opus 5.5 (D10) Share of time manipulating, five definitionsJoints moving81.1%End-effector moving66.0%Gripper engaged0.0%Contact force > 5 N79.3%Gripper engaged or in contact79.3% Gemini 4 Argon (D10) Share of time manipulating, five definitionsJoints moving75.6%End-effector moving66.6%Gripper engaged0.0%Contact force > 5 N63.5%Gripper engaged or in contact63.5% GPT-6 Astra (D10) Share of time manipulating, five definitionsJoints moving84.6%End-effector moving68.7%Gripper engaged0.0%Contact force > 5 N79.9%Gripper engaged or in contact79.9% GR00T N1.7 Share of time manipulating, five definitionsJoints moving99.8%End-effector moving95.5%Gripper engaged0.0%Contact force > 5 N75.4%Gripper engaged or in contact75.4% MolmoAct 2 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving93.6%Gripper engaged0.0%Contact force > 5 N65.6%Gripper engaged or in contact65.6% Pi 0.5 Share of time manipulating, five definitionsJoints moving99.8%End-effector moving96.2%Gripper engaged0.0%Contact force > 5 N63.6%Gripper engaged or in contact63.6% Rho Share of time manipulating, five definitionsJoints moving99.9%End-effector moving97.4%Gripper engaged0.0%Contact force > 5 N74.0%Gripper engaged or in contact74.0%
- SPARC smoothness
- DiT-Flow Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. Claude Fable 5.2 (D10) Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. Claude Opus 5.5 (D10) Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. Gemini 4 Argon (D10) Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. GPT-6 Astra (D10) Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. GR00T N1.7 Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. MolmoAct 2 Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. Pi 0.5 Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015. Rho Spectral arc length of the motionHow directly the arm moved, read from the shape of its speed spectrum. Unitless measure ≤ -1; the closer to -1, the smoother.The measure ignores how large a movement is, so a large and a small one compare directly. Shown here as the mean over this environment's measured episodes.Cutoff10 HzAmplitude threshold5%Zero-padding16×Signalend-effector translationBalasubramanian et al., On the analysis of movement smoothness, J Neuroeng Rehabil 2015.
Episode
Held-out benchmark environment
0:000:18