BenchmarkPAW-GEN-10-Agentic
Environmente06Put tools in standing toolbox
- Episodes
- Claude Fable 5.2 30 Claude Opus 5.5 30 DiT-Flow 30 Gemini 4 Argon 30 GPT-6 Astra 30 GR00T N1.7 30 MolmoAct 2 30 Pi 0.5 30 Rho 30
- Total duration
- Claude Fable 5.2 87:44 Claude Opus 5.5 66:09 DiT-Flow 6:01 Gemini 4 Argon 58:18 GPT-6 Astra 25:41 GR00T N1.7 7:03 MolmoAct 2 10:10 Pi 0.5 10:37 Rho 2:58
- Mean duration
- Claude Fable 5.2 175.5 s Claude Opus 5.5 132.3 s DiT-Flow 12.0 s Gemini 4 Argon 116.6 s GPT-6 Astra 51.4 s GR00T N1.7 14.1 s MolmoAct 2 20.3 s Pi 0.5 21.2 s Rho 5.9 s
- Success
- Claude Fable 5.2 77% Claude Opus 5.5 3% DiT-Flow 0% Gemini 4 Argon 3% GPT-6 Astra 0% GR00T N1.7 0% MolmoAct 2 0% Pi 0.5 3% Rho 0%
- Manipulation density
- Claude Fable 5.2 Share of time manipulating, five definitionsJoints moving78.2%End-effector moving56.7%Gripper engaged96.6%Contact force > 5 N47.7%Gripper engaged or in contact98.1% Claude Opus 5.5 Share of time manipulating, five definitionsJoints moving81.3%End-effector moving64.2%Gripper engaged63.7%Contact force > 5 N88.7%Gripper engaged or in contact94.9% DiT-Flow Share of time manipulating, five definitionsJoints moving97.8%End-effector moving94.3%Gripper engaged28.9%Contact force > 5 N74.4%Gripper engaged or in contact83.1% Gemini 4 Argon Share of time manipulating, five definitionsJoints moving43.4%End-effector moving31.7%Gripper engaged67.8%Contact force > 5 N78.9%Gripper engaged or in contact90.7% GPT-6 Astra Share of time manipulating, five definitionsJoints moving57.8%End-effector moving40.9%Gripper engaged30.7%Contact force > 5 N38.7%Gripper engaged or in contact60.6% GR00T N1.7 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving98.8%Gripper engaged31.4%Contact force > 5 N79.9%Gripper engaged or in contact87.3% MolmoAct 2 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving97.3%Gripper engaged31.3%Contact force > 5 N90.5%Gripper engaged or in contact93.0% Pi 0.5 Share of time manipulating, five definitionsJoints moving100.0%End-effector moving99.3%Gripper engaged38.8%Contact force > 5 N81.7%Gripper engaged or in contact86.9% Rho Share of time manipulating, five definitionsJoints moving100.0%End-effector moving98.3%Gripper engaged13.1%Contact force > 5 N72.8%Gripper engaged or in contact76.8%
- SPARC smoothness
- 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. 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. 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
0:003:25