# Playground Benchmark ## MuJoCo Playground PPO Benchmark Results SLM Lab v5.3 validates PPO on [MuJoCo Playground](https://google-deepmind.github.io/mujoco_playground/) — Google DeepMind's GPU-accelerated simulation platform. MuJoCo Playground uses the MJWarp backend (Warp-accelerated MJX) for physics, enabling massively parallel training with 2048 environments on GPU. SLM Lab wraps Playground environments as `gymnasium.VectorEnv` with DLPack zero-copy JAX→PyTorch transfer. All 54 environments use the `playground/` prefix in specs. Results below are from March 2026 benchmark runs. All trained models and metrics are publicly available on [HuggingFace](https://huggingface.co/datasets/SLM-Lab/benchmark). ### Methodology Results show **Trial-level** performance: 1. **Trial** = 4 Sessions with different random seeds 2. **Session** = One complete training run 3. **Score** = Final 100-checkpoint moving average (`total_reward_ma`) The trial score is the mean across 4 sessions. ### Standardized Settings | Category | num\_envs | max\_frame | log\_frequency | | ---------- | --------- | ---------- | -------------- | | Playground | 2048 | 100e6 | 10000 | ### Spec File **Spec file**: [ppo\_playground.yaml](https://github.com/kengz/SLM-Lab/blob/master/slm_lab/spec/benchmark_arc/ppo/ppo_playground.yaml) — all envs via `-s env=playground/ENV` ### Running Benchmarks ```bash # Requires: uv sync --group playground (JAX + MuJoCo Playground + MJWarp) source .env && slm-lab run-remote --gpu \ slm_lab/spec/benchmark_arc/ppo/ppo_playground.yaml SPEC_NAME train \ -s env=playground/ENV -s max_frame=100000000 -n NAME ``` ### Installation ```bash uv sync --group playground ``` This adds JAX, MuJoCo Playground, and MJWarp dependencies. Requires a CUDA GPU. *** ## Phase 5.1: DM Control Suite (25 envs) Classic control and locomotion tasks from the DeepMind Control Suite, ported to MJWarp GPU simulation. | ENV | MA | SPEC\_NAME | HF Data | | -------------------------------- | ------ | --------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | playground/AcrobotSwingup | 253.24 | ppo\_playground\_vnorm | [ppo\_playground\_acrobotswingup\_2026\_03\_12\_175809](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_acrobotswingup_2026_03_12_175809) | | playground/AcrobotSwingupSparse | 146.98 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_acrobotswingupsparse\_2026\_03\_14\_161212](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_acrobotswingupsparse_2026_03_14_161212) | | playground/BallInCup | 942.44 | ppo\_playground\_vnorm | [ppo\_playground\_ballincup\_2026\_03\_12\_105443](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_ballincup_2026_03_12_105443) | | playground/CartpoleBalance | 968.23 | ppo\_playground\_vnorm | [ppo\_playground\_cartpolebalance\_2026\_03\_12\_141924](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_cartpolebalance_2026_03_12_141924) | | playground/CartpoleBalanceSparse | 995.34 | ppo\_playground\_constlr | [ppo\_playground\_constlr\_cartpolebalancesparse\_2026\_03\_14\_000352](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_constlr_cartpolebalancesparse_2026_03_14_000352) | | playground/CartpoleSwingup | 729.09 | ppo\_playground\_constlr | [ppo\_playground\_constlr\_cartpoleswingup\_2026\_03\_17\_041102](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_constlr_cartpoleswingup_2026_03_17_041102) | | playground/CartpoleSwingupSparse | 521.98 | ppo\_playground\_constlr | [ppo\_playground\_constlr\_cartpoleswingupsparse\_2026\_03\_13\_233449](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_constlr_cartpoleswingupsparse_2026_03_13_233449) | | playground/CheetahRun | 883.44 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_cheetahrun\_2026\_03\_14\_161211](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_cheetahrun_2026_03_14_161211) | | playground/FingerSpin | 713.35 | ppo\_playground\_fingerspin | [ppo\_playground\_fingerspin\_fingerspin\_2026\_03\_13\_033911](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_fingerspin_fingerspin_2026_03_13_033911) | | playground/FingerTurnEasy | 663.58 | ppo\_playground\_vnorm | [ppo\_playground\_fingerturneasy\_2026\_03\_12\_175835](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_fingerturneasy_2026_03_12_175835) | | playground/FingerTurnHard | 590.43 | ppo\_playground\_vnorm\_constlr | [ppo\_playground\_vnorm\_constlr\_fingerturnhard\_2026\_03\_16\_234509](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_constlr_fingerturnhard_2026_03_16_234509) | | playground/FishSwim | 580.57 | ppo\_playground\_vnorm\_constlr\_clip03 | [ppo\_playground\_vnorm\_constlr\_clip03\_fishswim\_2026\_03\_14\_002112](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_constlr_clip03_fishswim_2026_03_14_002112) | | playground/HopperHop | 22.00 | ppo\_playground\_vnorm | [ppo\_playground\_hopperhop\_2026\_03\_12\_110855](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_hopperhop_2026_03_12_110855) | | playground/HopperStand | 237.15 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_hopperstand\_2026\_03\_14\_095438](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_hopperstand_2026_03_14_095438) | | playground/HumanoidRun | 18.83 | ppo\_playground\_humanoid | [ppo\_playground\_humanoid\_humanoidrun\_2026\_03\_14\_115522](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_humanoid_humanoidrun_2026_03_14_115522) | | playground/HumanoidStand | 114.86 | ppo\_playground\_humanoid | [ppo\_playground\_humanoid\_humanoidstand\_2026\_03\_14\_115516](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_humanoid_humanoidstand_2026_03_14_115516) | | playground/HumanoidWalk | 47.01 | ppo\_playground\_humanoid | [ppo\_playground\_humanoid\_humanoidwalk\_2026\_03\_14\_172235](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_humanoid_humanoidwalk_2026_03_14_172235) | | playground/PendulumSwingup | 637.46 | ppo\_playground\_pendulum | [ppo\_playground\_pendulum\_pendulumswingup\_2026\_03\_13\_033818](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_pendulum_pendulumswingup_2026_03_13_033818) | | playground/PointMass | 868.09 | ppo\_playground\_vnorm\_constlr | [ppo\_playground\_vnorm\_constlr\_pointmass\_2026\_03\_14\_095452](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_constlr_pointmass_2026_03_14_095452) | | playground/ReacherEasy | 955.08 | ppo\_playground\_vnorm | [ppo\_playground\_reachereasy\_2026\_03\_12\_122115](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_reachereasy_2026_03_12_122115) | | playground/ReacherHard | 946.99 | ppo\_playground\_vnorm | [ppo\_playground\_reacherhard\_2026\_03\_12\_123226](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_reacherhard_2026_03_12_123226) | | playground/SwimmerSwimmer6 | 591.13 | ppo\_playground\_vnorm\_constlr | [ppo\_playground\_vnorm\_constlr\_swimmerswimmer6\_2026\_03\_14\_000406](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_constlr_swimmerswimmer6_2026_03_14_000406) | | playground/WalkerRun | 759.71 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_walkerrun\_2026\_03\_14\_161354](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_walkerrun_2026_03_14_161354) | | playground/WalkerStand | 948.35 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_walkerstand\_2026\_03\_14\_161415](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_walkerstand_2026_03_14_161415) | | playground/WalkerWalk | 945.31 | ppo\_playground\_vnorm | [ppo\_playground\_vnorm\_walkerwalk\_2026\_03\_14\_161338](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_vnorm_walkerwalk_2026_03_14_161338) | | | | | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ![AcrobotSwingup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/AcrobotSwingup_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![AcrobotSwingupSparse](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/AcrobotSwingupSparse_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![BallInCup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/BallInCup_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![CartpoleBalance](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/CartpoleBalance_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![CartpoleBalanceSparse](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/CartpoleBalanceSparse_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![CartpoleSwingup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/CartpoleSwingup_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![CartpoleSwingupSparse](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/CartpoleSwingupSparse_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![CheetahRun](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/CheetahRun_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![FingerSpin](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/FingerSpin_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![FingerTurnEasy](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/FingerTurnEasy_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![FingerTurnHard](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/FingerTurnHard_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![FishSwim](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/FishSwim_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![HopperHop](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/HopperHop_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![HopperStand](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/HopperStand_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![HumanoidRun](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/HumanoidRun_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![HumanoidStand](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/HumanoidStand_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![HumanoidWalk](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/HumanoidWalk_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![PendulumSwingup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PendulumSwingup_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![PointMass](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PointMass_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![ReacherEasy](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/ReacherEasy_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![ReacherHard](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/ReacherHard_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![SwimmerSwimmer6](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/SwimmerSwimmer6_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![WalkerRun](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/WalkerRun_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![WalkerStand](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/WalkerStand_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![WalkerWalk](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/WalkerWalk_multi_trial_graph_mean_returns_ma_vs_frames.png) | | | *** ## Phase 5.2: Locomotion Robots (19 envs) Real-world robot locomotion — quadrupeds (Go1, Spot, Barkour) and humanoids (H1, G1, T1, Op3, Apollo, BerkeleyHumanoid) on flat and rough terrain. | ENV | MA | SPEC\_NAME | HF Data | | ----------------------------------------------- | ----- | ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | playground/ApolloJoystickFlatTerrain | 17.44 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_apollojoystickflatterrain\_2026\_03\_14\_210939](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_apollojoystickflatterrain_2026_03_14_210939) | | playground/BarkourJoystick | 0.0 | ppo\_playground\_loco | [ppo\_playground\_loco\_barkourjoystick\_2026\_03\_14\_194525](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_barkourjoystick_2026_03_14_194525) | | playground/BerkeleyHumanoidJoystickFlatTerrain | 32.29 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_berkeleyhumanoidjoystickflatterrain\_2026\_03\_14\_213019](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_berkeleyhumanoidjoystickflatterrain_2026_03_14_213019) | | playground/BerkeleyHumanoidJoystickRoughTerrain | 21.25 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_berkeleyhumanoidjoystickroughterrain\_2026\_03\_15\_150211](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_berkeleyhumanoidjoystickroughterrain_2026_03_15_150211) | | playground/G1JoystickFlatTerrain | 1.85 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_g1joystickflatterrain\_2026\_03\_15\_150219](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_g1joystickflatterrain_2026_03_15_150219) | | playground/G1JoystickRoughTerrain | -2.75 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_g1joystickroughterrain\_2026\_03\_19\_015137](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_g1joystickroughterrain_2026_03_19_015137) | | playground/Go1Footstand | 23.48 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_go1footstand\_2026\_03\_16\_174009](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_go1footstand_2026_03_16_174009) | | playground/Go1Getup | 18.16 | ppo\_playground\_loco\_go1 | [ppo\_playground\_loco\_go1\_go1getup\_2026\_03\_16\_132801](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_go1_go1getup_2026_03_16_132801) | | playground/Go1Handstand | 17.88 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_go1handstand\_2026\_03\_16\_155437](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_go1handstand_2026_03_16_155437) | | playground/Go1JoystickFlatTerrain | 0.0 | ppo\_playground\_loco | [ppo\_playground\_loco\_go1joystickflatterrain\_2026\_03\_14\_204658](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_go1joystickflatterrain_2026_03_14_204658) | | playground/Go1JoystickRoughTerrain | 0.00 | ppo\_playground\_loco | [ppo\_playground\_loco\_go1joystickroughterrain\_2026\_03\_15\_150321](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_go1joystickroughterrain_2026_03_15_150321) | | playground/H1InplaceGaitTracking | 11.95 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_h1inplacegaittracking\_2026\_03\_16\_170327](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_h1inplacegaittracking_2026_03_16_170327) | | playground/H1JoystickGaitTracking | 31.11 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_h1joystickgaittracking\_2026\_03\_16\_170412](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_h1joystickgaittracking_2026_03_16_170412) | | playground/Op3Joystick | 0.00 | ppo\_playground\_loco | [ppo\_playground\_loco\_op3joystick\_2026\_03\_15\_150120](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_op3joystick_2026_03_15_150120) | | playground/SpotFlatTerrainJoystick | 48.58 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_spotflatterrainjoystick\_2026\_03\_16\_180747](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_spotflatterrainjoystick_2026_03_16_180747) | | playground/SpotGetup | 19.39 | ppo\_playground\_loco | [ppo\_playground\_loco\_spotgetup\_2026\_03\_14\_213703](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_spotgetup_2026_03_14_213703) | | playground/SpotJoystickGaitTracking | 36.90 | ppo\_playground\_loco | [ppo\_playground\_loco\_spotjoystickgaittracking\_2026\_03\_19\_015106](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_spotjoystickgaittracking_2026_03_19_015106) | | playground/T1JoystickFlatTerrain | 13.42 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_t1joystickflatterrain\_2026\_03\_14\_220250](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_t1joystickflatterrain_2026_03_14_220250) | | playground/T1JoystickRoughTerrain | 2.58 | ppo\_playground\_loco\_precise | [ppo\_playground\_loco\_precise\_t1joystickroughterrain\_2026\_03\_15\_162332](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_precise_t1joystickroughterrain_2026_03_15_162332) | | | | | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ![ApolloJoystickFlatTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/ApolloJoystickFlatTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![BarkourJoystick](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/BarkourJoystick_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![BerkeleyHumanoidJoystickFlatTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/BerkeleyHumanoidJoystickFlatTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![G1JoystickFlatTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/G1JoystickFlatTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![Go1Footstand](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Go1Footstand_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![Go1Handstand](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Go1Handstand_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![H1InplaceGaitTracking](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/H1InplaceGaitTracking_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![H1JoystickGaitTracking](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/H1JoystickGaitTracking_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![Op3Joystick](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Op3Joystick_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![SpotFlatTerrainJoystick](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/SpotFlatTerrainJoystick_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![SpotGetup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/SpotGetup_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![SpotJoystickGaitTracking](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/SpotJoystickGaitTracking_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![BerkeleyHumanoidJoystickRoughTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/BerkeleyHumanoidJoystickRoughTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![Go1Getup](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Go1Getup_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![Go1JoystickFlatTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Go1JoystickFlatTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![Go1JoystickRoughTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/Go1JoystickRoughTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![T1JoystickFlatTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/T1JoystickFlatTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![T1JoystickRoughTerrain](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/T1JoystickRoughTerrain_multi_trial_graph_mean_returns_ma_vs_frames.png) | *** ## Phase 5.3: Manipulation (10 envs) Robotic manipulation — Panda arm pick/place, Aloha bimanual, Leap dexterous hand, and AeroCube orientation tasks. | ENV | MA | SPEC\_NAME | HF Data | | ----------------------------------- | -------- | ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | playground/AeroCubeRotateZAxis | -3.09 | ppo\_playground\_loco | [ppo\_playground\_loco\_aerocuberotatezaxis\_2026\_03\_20\_012502](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_aerocuberotatezaxis_2026_03_20_012502) | | playground/AlohaHandOver | 3.65 | ppo\_playground\_loco | [ppo\_playground\_loco\_alohahandover\_2026\_03\_15\_023712](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_alohahandover_2026_03_15_023712) | | playground/AlohaSinglePegInsertion | 220.93 | ppo\_playground\_manip\_aloha\_peg | [ppo\_playground\_manip\_aloha\_peg\_alohasinglepeginsertion\_2026\_03\_17\_122613](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_manip_aloha_peg_alohasinglepeginsertion_2026_03_17_122613) | | playground/LeapCubeReorient | 74.68 | ppo\_playground\_loco | [ppo\_playground\_loco\_leapcubereorient\_2026\_03\_15\_150420](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_leapcubereorient_2026_03_15_150420) | | playground/LeapCubeRotateZAxis | 91.65 | ppo\_playground\_loco | [ppo\_playground\_loco\_leapcuberotatezaxis\_2026\_03\_15\_150334](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_leapcuberotatezaxis_2026_03_15_150334) | | playground/PandaOpenCabinet | 11081.51 | ppo\_playground\_loco | [ppo\_playground\_loco\_pandaopencabinet\_2026\_03\_15\_150318](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_pandaopencabinet_2026_03_15_150318) | | playground/PandaPickCube | 4586.13 | ppo\_playground\_loco | [ppo\_playground\_loco\_pandapickcube\_2026\_03\_15\_023744](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_pandapickcube_2026_03_15_023744) | | playground/PandaPickCubeCartesian | 10.58 | ppo\_playground\_loco | [ppo\_playground\_loco\_pandapickcubecartesian\_2026\_03\_15\_023810](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_pandapickcubecartesian_2026_03_15_023810) | | playground/PandaPickCubeOrientation | 4281.66 | ppo\_playground\_loco | [ppo\_playground\_loco\_pandapickcubeorientation\_2026\_03\_19\_015108](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_pandapickcubeorientation_2026_03_19_015108) | | playground/PandaRobotiqPushCube | 1.31 | ppo\_playground\_loco | [ppo\_playground\_loco\_pandarobotiqpushcube\_2026\_03\_15\_042131](https://huggingface.co/datasets/SLM-Lab/benchmark/tree/main/data/ppo_playground_loco_pandarobotiqpushcube_2026_03_15_042131) | | | | | | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ![AeroCubeRotateZAxis](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/AeroCubeRotateZAxis_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![AlohaHandOver](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/AlohaHandOver_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![AlohaSinglePegInsertion](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/AlohaSinglePegInsertion_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![LeapCubeReorient](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/LeapCubeReorient_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![LeapCubeRotateZAxis](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/LeapCubeRotateZAxis_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![PandaOpenCabinet](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PandaOpenCabinet_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![PandaPickCube](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PandaPickCube_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![PandaPickCubeCartesian](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PandaPickCubeCartesian_multi_trial_graph_mean_returns_ma_vs_frames.png) | ![PandaPickCubeOrientation](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PandaPickCubeOrientation_multi_trial_graph_mean_returns_ma_vs_frames.png) | | ![PandaRobotiqPushCube](https://huggingface.co/datasets/SLM-Lab/benchmark/resolve/main/docs/plots/PandaRobotiqPushCube_multi_trial_graph_mean_returns_ma_vs_frames.png) | | | --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://slm-lab.gitbook.io/slm-lab/benchmark-results/playground-benchmark.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.