Unified Embodied VLM Reasoning with Robotic Action via Autoregressive Discretized Pre-training

TL;DR

This paper introduces ERIQ, a benchmark for embodied reasoning in robotics, and FACT, a discrete action tokenizer, to improve the integration of reasoning and precise control in robotic manipulation.

Contribution

It presents ERIQ for systematic evaluation of embodied reasoning and proposes FACT to bridge reasoning and control, enabling better robotic manipulation performance.

Findings

ERIQ reveals a strong correlation between reasoning and generalization.

FACT improves trajectory fidelity in discrete control sequences.

GenieReasoner outperforms prior methods in real-world robotic tasks.

Abstract

General-purpose robotic systems operating in open-world environments must achieve both broad generalization and high-precision action execution, a combination that remains challenging for existing Vision-Language-Action (VLA) models. While large Vision-Language Models (VLMs) improve semantic generalization, insufficient embodied reasoning leads to brittle behavior, and conversely, strong reasoning alone is inadequate without precise control. To provide a decoupled and quantitative assessment of this bottleneck, we introduce Embodied Reasoning Intelligence Quotient (ERIQ), a large-scale embodied reasoning benchmark in robotic manipulation, comprising 6K+ question-answer pairs across four reasoning dimensions. By decoupling reasoning from execution, ERIQ enables systematic evaluation and reveals a strong positive correlation between embodied reasoning capability and end-to-end VLA generalization. To bridge the gap from reasoning to precise execution, we propose FACT, a flow-matching-based action tokenizer that converts continuous control into discrete sequences while preserving high-fidelity trajectory reconstruction. The resulting GenieReasoner jointly optimizes reasoning and action in a unified space, outperforming both continuous-action and prior discrete-action baselines in real-world tasks. Together, ERIQ and FACT provide a principled framework for diagnosing and overcoming the reasoning-precision trade-off, advancing robust, general-purpose robotic manipulation. Project page: https://geniereasoner.github.io/GenieReasoner/