← Leaderboard
ScaleVLA
VLAGalbot et al. · Nov 16, 2026
Technical report from Tsinghua, PKU, SJTU and Galbot on scaling a VLA to a 552B mixture-of-experts LLM backbone with an action DiT trained from scratch. It reports LIBERO plus partial LIBERO-PRO and RoboCasa results; its real-robot result is 9/10 in-domain on a RoboDojo-style plug-in-charger task.
- Type
- VLA
- Parameters
- 552B
- Backbone
- Frozen DeepSeek-V4.1-Flash (MoE) + GR00T N1.7-style action DiT
- Weights
- Unknown
- License
- –
- Commercial use
- Unclear
- Released
- Nov 16, 2026
- Robots
- Franka Panda (sim)
- SO-101
- –
- LeRobot
- –
- Inputs
- –
- VRAM
- –
Benchmark results
1 of 13 benchmarks- Paper / model card ↗Nov 2026
#47 of 7096.1%, self-reported by the authors, not a leaderboard entry
Not reported on 12 benchmarks
Wrong or missing number? Report an error