BenchmarkPAW-GEN-10
Environmente06Put tools in standing toolbox
- 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 14:13 Claude Fable 5.2 (D10) 59:02 Claude Opus 5.5 (D10) 43:30 Gemini 4 Argon (D10) 50:15 GPT-6 Astra (D10) 24:46 GR00T N1.7 11:53 MolmoAct 2 10:56 Pi 0.5 13:45 Rho 11:17
- Mean duration
- DiT-Flow 28.4 s Claude Fable 5.2 (D10) 118.1 s Claude Opus 5.5 (D10) 87.0 s Gemini 4 Argon (D10) 100.5 s GPT-6 Astra (D10) 49.5 s GR00T N1.7 23.8 s MolmoAct 2 21.9 s Pi 0.5 27.5 s Rho 22.6 s
- Success
- DiT-Flow 0% Claude Fable 5.2 (D10) 30% Claude Opus 5.5 (D10) 0% Gemini 4 Argon (D10) 0% GPT-6 Astra (D10) 0% GR00T N1.7 0% MolmoAct 2 0% Pi 0.5 0% Rho 0%
- Manipulation density
- DiT-Flow Share of time manipulating, five definitionsJoints moving99.9%End-effector moving98.7%Gripper engaged11.9%Contact force > 5 N79.2%Gripper engaged or in contact82.5% Claude Fable 5.2 (D10) Share of time manipulating, five definitionsJoints moving77.7%End-effector moving56.0%Gripper engaged87.2%Contact force > 5 N46.7%Gripper engaged or in contact92.6% Claude Opus 5.5 (D10) Share of time manipulating, five definitionsJoints moving86.0%End-effector moving68.2%Gripper engaged37.2%Contact force > 5 N89.3%Gripper engaged or in contact93.1% Gemini 4 Argon (D10) Share of time manipulating, five definitionsJoints moving33.8%End-effector moving23.8%Gripper engaged32.6%Contact force > 5 N64.8%Gripper engaged or in contact71.3% GPT-6 Astra (D10) Share of time manipulating, five definitionsJoints moving60.2%End-effector moving41.3%Gripper engaged14.0%Contact force > 5 N52.3%Gripper engaged or in contact62.8% GR00T N1.7 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving98.6%Gripper engaged32.3%Contact force > 5 N73.9%Gripper engaged or in contact82.0% MolmoAct 2 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving96.2%Gripper engaged50.1%Contact force > 5 N71.0%Gripper engaged or in contact87.3% Pi 0.5 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.4%Gripper engaged74.9%Contact force > 5 N68.8%Gripper engaged or in contact92.6% Rho Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.0%Gripper engaged32.2%Contact force > 5 N74.0%Gripper engaged or in contact80.0%
- 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
0:000:13