π0.5
π0.5 places in the top three on 2 benchmarks.
12
Benchmarks with a score
0
Ranked first, out of five or more models
2
In the top three
Physical Intelligence Apr 2025
π0.5
Open weights Flagship
No Epoch Capabilities Index score yet
- RoboColiseum 57.4% #5 /9
- VLA-Arena 44.3% #5 /8
- RoboDojo 11.4% #2 /6
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| RoboLab | 28% | #1 of 4 | Official leaderboard | |
| RoboDojo | 11.4% | #2 of 6 | Official leaderboard | |
| LIBERO-PRO | 53% | #1 of 3 | Official leaderboard | |
| RoboArena | 1,608 | #1 of 3 | Official leaderboard | |
| RoboChallenge | 42.7% | #1 of 3 | Official leaderboard | |
| LIBERO | 96.8% | #3 of 6 | Model card | |
| PhAIL | 75.1% | #1 of 2 | Official leaderboard | |
| RoboTwin 2.0 | 58.3% | #1 of 2 | Official leaderboard | |
| RoboColiseum | 57.4% | #5 of 9 | Official leaderboard | |
| VLA-Arena | 44.3% | #5 of 8 | Official leaderboard | |
| RoboCasa365 | 16.9% | #3 of 4 | Official leaderboard | |
| PhAIL DROID: Everyday tasks | 12% | #3 of 3 | Official leaderboard |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.
Its system card
Every result in the longer documents the lab published with it.
Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better 28 May 2025