BenchmarkPAW-GEN-10
Environmente10Clamp electrical component
- 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 12:20 Claude Fable 5.2 (D10) 100:38 Claude Opus 5.5 (D10) 91:55 Gemini 4 Argon (D10) 139:28 GPT-6 Astra (D10) 47:51 GR00T N1.7 8:24 MolmoAct 2 18:10 Pi 0.5 14:15 Rho 22:37
- Mean duration
- DiT-Flow 24.7 s Claude Fable 5.2 (D10) 201.3 s Claude Opus 5.5 (D10) 183.8 s Gemini 4 Argon (D10) 278.9 s GPT-6 Astra (D10) 95.7 s GR00T N1.7 16.8 s MolmoAct 2 36.3 s Pi 0.5 28.5 s Rho 45.2 s
- Success
- DiT-Flow 23% Claude Fable 5.2 (D10) 87% Claude Opus 5.5 (D10) 33% Gemini 4 Argon (D10) 7% GPT-6 Astra (D10) 7% GR00T N1.7 0% MolmoAct 2 43% Pi 0.5 7% Rho 3%
- Manipulation density
- DiT-Flow Share of time manipulating, five definitionsJoints moving87.2%End-effector moving70.6%Gripper engaged82.1%Contact force > 5 N72.7%Gripper engaged or in contact93.3% Claude Fable 5.2 (D10) Share of time manipulating, five definitionsJoints moving88.2%End-effector moving71.1%Gripper engaged98.1%Contact force > 5 N52.7%Gripper engaged or in contact98.1% Claude Opus 5.5 (D10) Share of time manipulating, five definitionsJoints moving91.0%End-effector moving75.5%Gripper engaged98.6%Contact force > 5 N75.8%Gripper engaged or in contact98.6% Gemini 4 Argon (D10) Share of time manipulating, five definitionsJoints moving36.3%End-effector moving25.6%Gripper engaged17.6%Contact force > 5 N52.0%Gripper engaged or in contact56.3% GPT-6 Astra (D10) Share of time manipulating, five definitionsJoints moving74.5%End-effector moving54.7%Gripper engaged72.2%Contact force > 5 N93.8%Gripper engaged or in contact95.9% GR00T N1.7 Share of time manipulating, five definitionsJoints moving95.8%End-effector moving93.5%Gripper engaged80.4%Contact force > 5 N83.5%Gripper engaged or in contact92.5% MolmoAct 2 Share of time manipulating, five definitionsJoints moving99.3%End-effector moving91.7%Gripper engaged91.3%Contact force > 5 N86.8%Gripper engaged or in contact97.1% Pi 0.5 Share of time manipulating, five definitionsJoints moving96.7%End-effector moving85.9%Gripper engaged88.6%Contact force > 5 N85.9%Gripper engaged or in contact96.1% Rho Share of time manipulating, five definitionsJoints moving99.3%End-effector moving96.0%Gripper engaged93.1%Contact force > 5 N86.6%Gripper engaged or in contact97.1%
- 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
0:000:29