Alibaba’s RynnBrain 1.1 scales embodied foundation models up to 122B-A10B
Alibaba-DAMO-Academy · hf · 2026-07-21
What it is
RynnBrain 1.1 is Alibaba DAMO Academy’s embodied foundation model family, spanning 2B, 9B, and 122B-A10B scales.
What changed from v1.0
- Built with a unified spatio-temporal and physically grounded training framework.
- Covers embodied perception, spatial reasoning, localization, and planning.
- Adds contact-point prediction across the family.
- Adds native 3D grounding for the 2B and 9B models, making outputs more directly aligned with manipulation.
- RynnBrain-VLA introduces a unified cross-embodiment action space plus embodiment-specific masking.
Deployment and results
- Deployed on Unitree G1, Astribot-S1, and Tianji-Wuji.
- The 122B-A10B model reportedly beats all evaluated proprietary and open-source models on VSI-Bench, MMSI, and RefSpatial-Bench.
- Real-robot experiments show RynnBrain-initialized policies outperform Qwen-based and representative generalist VLAs, and joint multi-task / multi-embodiment training improves success rates and process scores.
More from Embodied
- Bedrock Robotics Raises $600M Series C as Autonomous Excavators Do Paid Work on US Job Sites — davidpattersonx · 2026-10-27
- UniAction: unified action conditioning lets one world model handle mixed action types — ZeYanjie · 2026-10-27
- StackChan robot performs an AI-generated skit, orchestrated by Hermes Agent via JSON script — alexcovo_eth · 2026-10-27
- DONIC hits 86.79% on bimanual dexterous tracking, crushing prior SOTA's 9.43% — liuziwei7 · 2026-10-27
- Bedrock Robotics raises $600M as its fully autonomous excavators do paid work on live job sites — Sethwinterroth · 2026-10-27
- AW-0 debuts: one generative world action model plans in code, predicts, and controls robots — trevordarrell · 2026-10-27