HY-Embodied-0.5 develops 2B and 32B embodied foundation models with MoT architecture and self-evolving post-training that outperform peers on 16 of 22 benchmarks and enable effective real-world robot control via VLA fine-tuning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
HY-Embodied-0.5: Embodied Foundation Models for Real-World Agents
HY-Embodied-0.5 develops 2B and 32B embodied foundation models with MoT architecture and self-evolving post-training that outperform peers on 16 of 22 benchmarks and enable effective real-world robot control via VLA fine-tuning.