BenchmarkPAW-GEN-10-Agentic
Environmente08Place spray bottles upright, nozzle forward (bimanual handover)
- Episodes
- Gemini 4 Argon 30 Claude Fable 5.2 30 Claude Opus 5.5 30 DiT-Flow 30 GPT-6 Astra 30 GR00T N1.7 30 MolmoAct 2 30 Pi 0.5 30 Rho 30
- Total duration
- Gemini 4 Argon 54:01 Claude Fable 5.2 44:59 Claude Opus 5.5 61:31 DiT-Flow 10:21 GPT-6 Astra 103:04 GR00T N1.7 19:56 MolmoAct 2 19:36 Pi 0.5 9:01 Rho 25:47
- Mean duration
- Gemini 4 Argon 108.0 s Claude Fable 5.2 90.0 s Claude Opus 5.5 123.0 s DiT-Flow 20.7 s GPT-6 Astra 206.1 s GR00T N1.7 39.9 s MolmoAct 2 39.2 s Pi 0.5 18.0 s Rho 51.6 s
- Success
- Gemini 4 Argon 90% Claude Fable 5.2 100% Claude Opus 5.5 90% DiT-Flow 0% GPT-6 Astra 97% GR00T N1.7 10% MolmoAct 2 0% Pi 0.5 10% Rho 20%
- Manipulation density
- Gemini 4 Argon Share of time manipulating, five definitionsJoints moving86.8%End-effector moving68.8%Gripper engaged97.5%Contact force > 5 N40.3%Gripper engaged or in contact97.5% Claude Fable 5.2 Share of time manipulating, five definitionsJoints moving95.2%End-effector moving79.9%Gripper engaged88.4%Contact force > 5 N65.7%Gripper engaged or in contact94.8% Claude Opus 5.5 Share of time manipulating, five definitionsJoints moving89.1%End-effector moving78.6%Gripper engaged83.8%Contact force > 5 N81.2%Gripper engaged or in contact91.3% DiT-Flow Share of time manipulating, five definitionsJoints moving98.7%End-effector moving92.4%Gripper engaged40.8%Contact force > 5 N55.8%Gripper engaged or in contact68.6% GPT-6 Astra Share of time manipulating, five definitionsJoints moving81.2%End-effector moving53.8%Gripper engaged68.3%Contact force > 5 N46.7%Gripper engaged or in contact80.7% GR00T N1.7 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.5%Gripper engaged68.8%Contact force > 5 N68.6%Gripper engaged or in contact89.2% MolmoAct 2 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.6%Gripper engaged68.6%Contact force > 5 N76.1%Gripper engaged or in contact91.0% Pi 0.5 Share of time manipulating, five definitionsJoints moving99.9%End-effector moving98.4%Gripper engaged53.6%Contact force > 5 N69.5%Gripper engaged or in contact83.6% Rho Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.0%Gripper engaged68.9%Contact force > 5 N65.7%Gripper engaged or in contact91.2%
- SPARC smoothness
- Gemini 4 Argon 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 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 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. GPT-6 Astra 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
0:001:53