Paper page - Kairos: A Native World Model Stack for Physical AI

Source: https://huggingface.co/papers/2606.16533 Published on Jun 16

#2 Paper of the day Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

Kairos is a native world model framework that learns from diverse experiences, maintains persistent states through hybrid temporal attention, and supports efficient deployment for physical AI applications.

World modelsare transitioning from passive visual generators to foundational, operational infrastructure for Physical AI: they must natively acquire world knowledge from heterogeneous experience, maintain persistent states over long horizons, and execute efficiently within real deployment constraints. We introduce Kairos, a native world model stack designed around these requirements. (1) Kairos learns the world by pioneering aNative Pre-training Paradigmgoverned by aCross-Embodiment Data Curriculum, which organizes open-world videos, human behavioral data, and robot interactions into a progressive developmental pathway. (2) Kairos maintains the world by unified world understanding, generation, and prediction within aNative Unified Architectureequipped withHybrid Linear Temporal Attention, wheresliding-window attentioncaptures local dynamics,dilated sliding windowscapture mid-range dependencies, andgated linear attentionmaintains persistent global memory. We establish formal theoretical bounds demonstrating that thistemporal factorizationstrictly limitserror accumulation, mathematically guaranteeing state propagation across extended horizons. (3) Kairos runs the world by incorporating aDeployment-Aware System Co-Designto support low-latency rollout generation on server and consumer-grade hardware for real-world observation-action-feedback loops. Experiments onembodied world-model, long-horizon, andaction-policy benchmarksshow that Kairos achieves top level performance while offering a strong efficiency-capability trade-off. Together, these results position Kairos as a cohesive operational foundation for future self-evolving physical intelligence.

View arXiv pageView PDFGitHub712Add to collection

Get this paper in your agent:

hf papers read 2606\.16533

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.16533 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.16533 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.16533 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.