OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation

2026-04-20

AI Summary

Existing implicit chain-of-thought (CoT) approaches in vision-language-action (VLA) autonomous driving suffer from performance limitations due to their reliance solely on compressed linguistic representations while neglecting the causal dynamics of driving scenes. In contrast, explicit CoT methods incur prohibitively high autoregressive inference latency, hindering real-time deployment. This work proposes OneVL, a novel framework that employs two auxiliary decoders—a language decoder reconstructing textual CoT and a visual world-model decoder predicting future frame tokens—to jointly supervise latent state learning. This design internalizes the causal structure underlying road geometry, agent motion, and environmental changes. Through a three-stage training strategy aligning trajectories, language, and visual objectives, OneVL achieves, for the first time, simultaneous improvements in both accuracy and inference speed over explicit CoT, attaining state-of-the-art performance across four benchmarks with latency comparable to direct answer prediction methods.

Technology Category

• Computer Vision: Language and Vision
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Chain-of-Thought (CoT) reasoning has become a powerful driver of trajectory prediction in VLA-based autonomous driving, yet its autoregressive nature imposes a latency cost that is prohibitive for real-time deployment. Latent CoT methods attempt to close this gap by compressing reasoning into continuous hidden states, but consistently fall short of their explicit counterparts. We suggest that this is due to purely linguistic latent representations compressing a symbolic abstraction of the world, rather than the causal dynamics that actually govern driving. Thus, we present OneVL (One-step latent reasoning and planning with Vision-Language explanations), a unified VLA and World Model framework that routes reasoning through compact latent tokens supervised by dual auxiliary decoders. Alongside a language decoder that reconstructs text CoT, we introduce a visual world model decoder that predicts future-frame tokens, forcing the latent space to internalize the causal dynamics of road geometry, agent motion, and environmental change. A three-stage training pipeline progressively aligns these latents with trajectory, language, and visual objectives, ensuring stable joint optimization. At inference, the auxiliary decoders are discarded and all latent tokens are prefilled in a single parallel pass, matching the speed of answer-only prediction. Across four benchmarks, OneVL becomes the first latent CoT method to surpass explicit CoT, delivering state-of-the-art accuracy at answer-only latency, and providing direct evidence that tighter compression, when guided in both language and world-model supervision, produces more generalizable representations than verbose token-by-token reasoning. Project Page: https://xiaomi-embodied-intelligence.github.io/OneVL

Problem

• Chain-of-Thought reasoning
• autonomous driving
• latency
• latent representation
• causal dynamics

Innovation

• Latent Chain-of-Thought
• Vision-Language Action (VLA)
• World Model
• Causal Dynamics
• Parallel Inference