← Leaderboard
Galaxea G0.5
VLAGalaxea AI · Jun 1, 2026 · built on GalaxeaVLA (G0)
Successor to G0 that combines reasoning and action in one autoregressive stream. It is the strongest raw policy on SO-101 Bench, and its DROID variant is ranked on RoboArena.
- Type
- VLA
- Parameters
- —
- Backbone
- Qwen3.5-2B
- Weights
- Gated
- License
- G0.5 Community License
- Commercial use
- Not allowed
- Released
- Jun 1, 2026
- Robots
- SO-101, DROID Franka, Galaxea R1 Lite, Galaxea R1 Pro
- SO-101
- Yes
- LeRobot
- –
Benchmark results
6 of 13 benchmarksRoboDojo (sim)Simulation
- Official leaderboard ↗Aug 4, 2026
#24 of 4014.9%
- 93.3%, self-reported by the authors, not a leaderboard entryPaper / model card ↗Aug 2026· self-reported, 50 clean + 500 randomized demos· non-standard
93.3%, self-reported by the authors, not a leaderboard entrynot ranked
- Paper / model card ↗Aug 2026
#10 of 4698.9%, self-reported by the authors, not a leaderboard entry
- Paper / model card ↗Aug 2026· visual matching
#5 of 1987.3%, self-reported by the authors, not a leaderboard entry
RoboArenaReal robot
- Official leaderboard ↗Nov 17, 2026· n = 45
#6 of 131559
SO-101 BenchLow-cost arm
- Paper / model card ↗Oct 27, 2026· n = 916
#4 of 1160.8%, self-reported by the authors, not a leaderboard entry
Not reported on 7 benchmarks
Adoption
updated Nov 17, 2026Speed per task
| Setting | Per task | Per planner call | Calls | Cost | Source |
|---|---|---|---|---|---|
| SO-101 real robot, 45 Sim2Real layouts | 38 s | – | – | – | Oct 27, 2026 ↗ |
Trained on
Wrong or missing number? Report an error