Liquid AI open-sources Open d1 decision models: d1-3B runs at 35ms per image on Jetson Thor
TheTuringPost · x · 2026-10-12
Liquid AI, known for on-device efficient models, released Open d1 decision models: d1-3B and d1-omni-600M, both making decisions from images, with the smaller one supporting experimental audio — all open-weight.
Liquid reports 35 ms per image for d1-3B on an NVIDIA Jetson AGX Thor, and the models were built with the target hardware in mind.
Implications for physical AI: robots could make many quick checks during a task without calling a big generative model each time; checks can run on-device without streaming camera frames to the cloud; Open d1 can monitor surroundings while other systems plan and move.
Related event: Liquid AI open-sources fast multimodal d1 decision models(3 posts)→
More from Embodied
- Unitree open-sources 6B humanoid model UnifoLM-WLA-1.0, claims fine-tuning on a 24GB GPU — lmoroney · 2026-10-12
- Tesla Cybercab tracker logs 145 on the road, 319 registered with Texas DMV — JOBhakdi · 2026-10-12
- Fleet Week traffic trapped Andrew Chen's Waymo for 15 min until humans-in-the-loop rerouted it — andrewchen · 2026-10-12
- COMODO distills video intelligence into IMU sensors for camera-free egocentric AI glasses — flosalim · 2026-10-12
- DirtyMoCap recovers 3D motion from noisy unordered markers, with 100x faster CUDA solver — jamestagg · 2026-10-12
- AI agent deploys a micro factory end-to-end: €20k cost, ROI under 2 years — ihorbeaver · 2026-10-12