← Leaderboard
EO-1
VLAShanghai AI Lab · Aug 28, 2025
Unified embodied foundation model trained with interleaved vision-text-action pretraining (Shanghai AI Lab with Fudan, AgiBot and NWPU), reported on LIBERO, SimplerEnv and several real robots including SO-100.
- Type
- VLA
- Parameters
- 3B
- Backbone
- Qwen2.5-VL-3B
- Weights
- Open
- License
- MIT
- Commercial use
- Allowed
- Released
- Aug 28, 2025
- Robots
- Franka Panda (sim), WidowX (Bridge), AgiBot G-1, Franka Panda, WidowX 250 S, SO-100
- SO-101
- Yes
- LeRobot
- –
Benchmark results
2 of 13 benchmarks- Paper / model card ↗Aug 2025
#22 of 4698.2%, self-reported by the authors, not a leaderboard entry
- Paper / model card ↗Aug 2025· visual matching
#8 of 1972.7%, self-reported by the authors, not a leaderboard entry
Not reported on 11 benchmarks
Adoption
updated Nov 17, 2026Running the model
| Hardware | Precision | Config | Latency / chunk | Steps | VRAM | Chunk | Control | Source |
|---|---|---|---|---|---|---|---|---|
| RTX 4090 | – | multi-view inference | – | 10 | 6 GB | 16 | – | Aug 28, 2025 ↗ |
Trained on
Wrong or missing number? Report an error