AetherAI's CausalWM tops TriWorldBench with causal chain-of-thought world modeling
机器之心 · wechat · 2026-09-20
AetherAI released CausalWM, a 16B causal world model introducing Causal Chain-of-Thought: instead of jumping straight to future frames, it explicitly reasons Observation → Motion (optical flow) → Geometry (depth/pointmap) → Future, with stage-ordered attention masks preventing answer leakage. Trained on 30k hours of embodied video in three stages (pixel pretraining on LTX-2.3-22B, Causal CoT mid-training, multi-objective RL), it tops TriWorldBench (TWB-Score 66.04) and PAI-Bench. The causal chain also doubles as a control interface for intervention-style queries. Paper, code, and weights are open.
More from Embodied
- Dev shows robot arm controlled by Claude, quips robotics is 'far too easy' and returns to GPU kernels — KuterDinel · 2026-09-20
- Sitzmann: agents may choose specialized models, so all bets are off again — vincesitzmann · 2026-09-20
- Sitzmann clarifies: only learned systems generalize to unknown in-the-wild objects — vincesitzmann · 2026-09-20
- Neuralink releases 'Speaking With The Mind', a 7-minute film on its BCI — Competitive_Travel16 · 2026-09-20
- GPT-6 Astra drives a $150 robot arm to paint, hits 95% on block pickup — 机器之心 · 2026-09-20
- LEXSUS robots grill live for 550K viewers, self-correcting failures to prove cross-embodiment AI — 创业邦 · 2026-09-20