Robbyant, Ant Group’s embodied-intelligence unit, has released LingBot-World-Infinity (LingBot-World 2.0), a causal video generation model that behaves as an interactive world simulator. It attacks two failure modes: long-horizon drift and interactive latency. The research team claims four upgrades over LingBot-World: an unbounded interaction horizon with consistent output quality, a distilled real-time variant sufficient to drive 720p video streams at 60 fps, a broader action space including attacking, archery, spell-casting, and shooting, and an agentic harness pairing a pilot agent with a director agent. The primary model is 14B, with a lightweight 1.3B counterpart described as deployable on a single GPU. The core contribution is the Mixture of Bidirectional and Autoregressive (MoBA) Attention Mask, which appends a bidirectional full-attention block to the teacher forcing mask to act as a regularizer. Pre-training optimizes a conditional flow-matching objective with rectified-flow interpolation. Post-training compresses the multi-step teacher into a few-step student using consistency distillation and distribution matching distillation (DMD), applied over long self-rollout trajectories to optimize the student on the state distribution its own predictions induce.