BenchmarkPAW-GEN-10
Environmente08Place spray bottles upright, nozzle forward (bimanual handover)
- Episodes
- Rho 30 Claude Fable 5.2 (D10) 30 Claude Opus 5.5 (D10) 30 DiT-Flow 30 Gemini 4 Argon (D10) 30 GPT-6 Astra (D10) 30 GR00T N1.7 30 MolmoAct 2 30 Pi 0.5 30
- Total duration
- Rho 26:13 Claude Fable 5.2 (D10) 42:10 Claude Opus 5.5 (D10) 51:18 DiT-Flow 32:17 Gemini 4 Argon (D10) 40:02 GPT-6 Astra (D10) 90:13 GR00T N1.7 20:00 MolmoAct 2 25:04 Pi 0.5 35:02
- Mean duration
- Rho 52.4 s Claude Fable 5.2 (D10) 84.3 s Claude Opus 5.5 (D10) 102.6 s DiT-Flow 64.6 s Gemini 4 Argon (D10) 80.1 s GPT-6 Astra (D10) 180.4 s GR00T N1.7 40.0 s MolmoAct 2 50.1 s Pi 0.5 70.1 s
- Success
- Rho 40% Claude Fable 5.2 (D10) 87% Claude Opus 5.5 (D10) 63% DiT-Flow 67% Gemini 4 Argon (D10) 60% GPT-6 Astra (D10) 70% GR00T N1.7 27% MolmoAct 2 67% Pi 0.5 43%
- Manipulation density
- Rho Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.5%Gripper engaged69.4%Contact force > 5 N66.7%Gripper engaged or in contact90.2% Claude Fable 5.2 (D10) Share of time manipulating, five definitionsJoints moving95.0%End-effector moving79.6%Gripper engaged87.5%Contact force > 5 N63.6%Gripper engaged or in contact94.9% Claude Opus 5.5 (D10) Share of time manipulating, five definitionsJoints moving88.8%End-effector moving78.8%Gripper engaged77.2%Contact force > 5 N81.6%Gripper engaged or in contact90.9% DiT-Flow Share of time manipulating, five definitionsJoints moving100.0%End-effector moving98.9%Gripper engaged70.3%Contact force > 5 N72.5%Gripper engaged or in contact94.5% Gemini 4 Argon (D10) Share of time manipulating, five definitionsJoints moving80.7%End-effector moving63.1%Gripper engaged98.0%Contact force > 5 N45.6%Gripper engaged or in contact98.1% GPT-6 Astra (D10) Share of time manipulating, five definitionsJoints moving75.7%End-effector moving50.9%Gripper engaged60.7%Contact force > 5 N46.8%Gripper engaged or in contact79.1% GR00T N1.7 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.7%Gripper engaged99.8%Contact force > 5 N71.6%Gripper engaged or in contact99.8% MolmoAct 2 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.5%Gripper engaged99.8%Contact force > 5 N68.0%Gripper engaged or in contact99.8% Pi 0.5 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.4%Gripper engaged99.9%Contact force > 5 N69.7%Gripper engaged or in contact99.9%
- SPARC smoothness
- 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. 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. 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. 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.
Episode
Held-out benchmark environment
0:000:19