Robotic Manipulation (Long) on LIBERO Long 1.0 (test)
89.3SRTrojanRobot
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TrojanRobotVLA Model=π02026.01 | 89.3 | 85.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State BackdoorVLA Model=π02026.01 | 89.2 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TrojanRobotVLA Model=SmolVLA2026.01 | 87.2 | 80.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NoneVLA Model=π02026.01 | 86.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State BackdoorVLA Model=SmolVLA2026.01 | 86.1 | 82.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NoneVLA Model=SmolVLA2026.01 | 85.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BadVLAVLA Model=π02026.01 | 84.6 | 84.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State BackdoorVLA Model=OpenVLA2026.01 | 84.1 | 89 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BadVLAVLA Model=OpenVLA2026.01 | 83.5 | 80.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TrojanRobotVLA Model=OpenVLA2026.01 | 82.4 | 84.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NoneVLA Model=OpenVLA2026.01 | 82.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NoneVLA Model=ACT2026.01 | 52.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NoneVLA Model=DP2026.01 | 49.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State BackdoorVLA Model=ACT2026.01 | 48.3 | 45.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TrojanRobotVLA Model=ACT2026.01 | 47.5 | 43.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| State BackdoorVLA Model=DP2026.01 | 46.1 | 43.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BadVLAVLA Model=ACT2026.01 | 44.8 | 46.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BadVLAVLA Model=DP2026.01 | 43.1 | 46.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TrojanRobotVLA Model=DP2026.01 | 42.2 | 24.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BadVLAVLA Model=SmolVLA2026.01 | 40.7 | 83.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| H-WM-guided π0.5guidance=H-WM, base_model=π0.52026.02 | — | — | — | 98 | 94 | 86.7 | 60 | 74 | 46 | 70.7 | 42 | 95 | 82 | 84.9 | 64.8 | |
| KosMos-2Framework=VLM4VLA, Size=1.7B2026.01 | — | — | 55 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Language-guided π0.5guidance=natural language task decomposition, base_model=π0.52026.02 | — | — | — | 84.7 | 54 | 80.7 | 42 | 68 | 24 | 41.3 | 4 | 59.5 | 10 | 66.8 | 26.8 | |
| Logic-guided π0.5guidance=logic-only, base_model=π0.52026.02 | — | — | — | 95.3 | 86 | 84.7 | 58 | 54.7 | 16 | 39.3 | 4 | 92 | 78 | 73.2 | 48.4 | |
| OpenVLA (Llama-2)Size=7.7B2026.01 | — | — | 53.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Paligemma-1Framework=VLM4VLA, Size=2.9B2026.01 | — | — | 44.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Paligemma-2Framework=VLM4VLA, Size=3.0B2026.01 | — | — | 46.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| pi0* (Paligemma-1)Size=3.1B2026.01 | — | — | 46 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5VL-3BFramework=VLM4VLA, Size=3.8B2026.01 | — | — | 43 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5VL-7BFramework=VLM4VLA, Size=8.3B2026.01 | — | — | 45 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3VL-2BFramework=VLM4VLA, Size=2.1B2026.01 | — | — | 55.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3VL-30B-A3BFramework=VLM4VLA, Size=31.1B2026.01 | — | — | 46.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3VL-4BFramework=VLM4VLA, Size=4.4B2026.01 | — | — | 44.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen3VL-8BFramework=VLM4VLA, Size=8.8B2026.01 | — | — | 46.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ThinkAct (Qwen2.5VL-7B)Size=7.4B2026.01 | — | — | 70.9 | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0guidance=none (end-to-end)2026.02 | — | — | — | 54 | 0 | 62 | 28 | 44 | 0 | 31.3 | 0 | 34 | 0 | 45.1 | 5.6 | |
| π0.5guidance=none (end-to-end)2026.02 | — | — | — | 66 | 4 | 73.3 | 24 | 54.7 | 4 | 44.7 | 0 | 38 | 0 | 55.3 | 6.4 |