Benchmark Almanac

SafeLIBERO

AEGIS 68.1% tops its leaderboard π0.5 57.8% leads the models we track
Not in model cards

VLA policies run 16 LIBERO tasks with an added obstacle at two levels, near the target or in the path. 50 episodes per scenario. Scored by collision avoidance and task success.

Introduced with the AEGIS safety layer and released December 20th, 2025. At least five 2026 papers report it, often with different metrics or dropped tasks, so scores rarely line up.

Released
December 2025
Kind
Simulated robots
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Last model added
April 2025
Scores refreshed
November 12th, 2026
3
Entries on its own leaderboard, 2 models we track

AEGIS tops the leaderboard with 68.1%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is π0.5, at 57.8%.

  1. 68.1 AEGIS
  2. 57.8 π0.5
  3. 26.2 OpenVLA-OFT
  1. 1 AEGIS translational action space 68.1%
  2. 2 π0.5 full action space 57.8%
  3. 3 OpenVLA-OFT full action space 26.2%

Source: Official SafeLIBERO leaderboard

#EntryScoreSettingDateSource
1 AEGIS 68.1% translational action space Official leaderboard
2 π0.5 57.8% full action space Apr 2025 Official leaderboard
3 OpenVLA-OFT 26.2% full action space Feb 2025 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Physical Intelligence
Own leaderboard Not tested since Dec 2025
π0.5 57.8%

Source: Official SafeLIBERO leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) π0.5 OpenVLA-OFT 0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) π0.5 OpenVLA-OFT

Source: Official SafeLIBERO leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Robotics

All 104 in Robotics
  • ERQA In use 400 multiple-choice visual questions from robot and egocentric footage on spatial, trajectory, action and state reasoning. Accuracy. Gemini Robotics ER 2 78.5%
  • RoboDojo In use Generalist manipulation policies run 42 Isaac Sim tasks and 18 remote real-robot tasks. Scored by task score and success rate. GPT-6 Astra 29%
  • RoboLab In use Zero-shot DROID Franka policies in Isaac Lab on 120 tabletop tasks at three instruction levels, 10 episodes each. Success rate. FLUX 3 Action 42.9%
  • Progress classification (Google DeepMind) In use Models assign each frame of a robot video to one of five task-progress bands from 0 to 100 percent. Scored by accuracy. Gemini Robotics ER 2 57.4%
  • Human proximity (Google DeepMind robotics) In use Models judge from robot camera views whether a human is within 1 meter, a spatial awareness check for safe robot behavior. Scored by accuracy. Gemini Robotics ER 2 93%
  • Moment finding (Google DeepMind) In use Models find the moment in a robot video when an event happens. Scored by accuracy and mean absolute time distance in seconds. Gemini Robotics ER 2 91.3%