TARS Unveils Embodied AI Model AWE 3.5 with Cross-Task Generalization

TARS released AWE 3.5, an embodied AI foundation model trained on a million hours of data, demonstrating cross-scene tasks like packing and cable insertion via an embodiment-native architecture with latent actions and geometric priors.

2026-08-18 ~ 2026-08-19 · 2 related posts

• TARS AWE 3.5 aims for generalization with embodied-native architecture — FellMentKE · 2026-08-18
• TARS AWE 3.5 Shows Cross-Scene Generalization with Million-Hour Training — heyshrutimishra · 2026-08-19