Tesla FSD v14.3.6 Revealed: 10B MoE Model and RL Enhancements
qinzytech · x · 2026-08-17
Details of Tesla FSD v14.3.6 have been revealed, featuring a 10B parameter Mixture-of-Experts (MoE) model with 2B active parameters (confirmed by Elon Musk). The vision encoder has improved training for rare scenes, and numerous Reinforcement Learning (RL) tricks have been applied. Observers note this is likely the most well-trained Vision-Language-Action (VLA) model to date.
More from Embodied
- AinaInterface team moves to factory to build next-gen HCI — pritisinghhhh · 2026-08-17
- AI Vision: Robots to handle wet lab experiments — dr_alphalyrae · 2026-08-17
- HopTo's Hop-1 Robot Uses Hopping to Cut Energy Use by 75% — MarwaEldiwiny · 2026-08-17
- New robot training method: taga-style attention enables massive parallelization — ChongZzZhang · 2026-08-17
- FCC adds foreign ground robots to Covered List; clarifies it's not a ban and targets production location — the-uncanny-squad · 2026-08-17
- AME2 neural mapping adapted for Livox lidar trained within 1 day — ChongZzZhang · 2026-08-17