VLANeXt Family: 500+ controlled experiments distill 12 practical recipes for VLA models
ccloy · x · 2026-09-30
Researchers released VLANeXt Family, a systematic study of Vision-Language-Action (VLA) models based on 500+ controlled experiments, distilling 12 practical recipes for building strong VLAs.
- Extends beyond the original VLANeXt into four emerging paradigms: model scaling, Latent Action Models (LAMs), JEPA-style predictive representation learning, and World Action Models (WAMs)
- The variants form a unified framework for studying both current and emerging VLA paradigms
- Paper, project page, and code are all publicly available
More from Embodied
- USC's CLAM learns robot policies from unlabeled videos, 2-3x success over baselines — ebiyik_ · 2026-09-30
- NVIDIA's Robo Olympics uses Codex and natural language to train robot skills in physics simulation — RevLebaredian · 2026-09-30
- A huggable humanoid robot called Baymax shows up at IROS 2026 — CyberRobooo · 2026-09-30
- Simple-WAM: One forward pass replaces video denoising, keeping world action models' generalization — Tsinghua-LeapLab · 2026-09-30
- Open-source humanoid Roboparty shows up at IROS, alongside a Gundam you can touch — 4310sy · 2026-09-30
- Scoble says Google Glass is coming back: 'We were early, not wrong' — Scobleizer · 2026-09-30