A minimal VLA baseline using a strong VLM backbone matches or exceeds more complex models across multiple robotics benchmarks including a 20% gain over π0.5 on real-world RoboChallenge.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
StarVLA-$\alpha$: Reducing Complexity in Vision-Language-Action Systems
A minimal VLA baseline using a strong VLM backbone matches or exceeds more complex models across multiple robotics benchmarks including a 20% gain over π0.5 on real-world RoboChallenge.