VIKI-R: Coordinating Embodied Multi-Agent Cooperation via Reinforcement Learning

TL;DR

This paper introduces VIKI-R, a reinforcement learning framework for embodied multi-agent cooperation, evaluated on the new VIKI-Bench benchmark, demonstrating improved coordination and visual reasoning across diverse robot types.

Contribution

The work presents VIKI-Bench, a hierarchical benchmark for embodied multi-agent cooperation, and VIKI-R, a novel RL-based method fine-tuning vision-language models for better multi-agent coordination.

Findings

VIKI-R outperforms baseline methods across all task levels.

Reinforcement learning fosters emergent compositional cooperation among heterogeneous agents.

VIKI-Bench provides a comprehensive platform for evaluating embodied multi-agent visual reasoning.

Abstract

Coordinating multiple embodied agents in dynamic environments remains a core challenge in artificial intelligence, requiring both perception-driven reasoning and scalable cooperation strategies. While recent works have leveraged large language models (LLMs) for multi-agent planning, a few have begun to explore vision-language models (VLMs) for visual reasoning. However, these VLM-based approaches remain limited in their support for diverse embodiment types. In this work, we introduce VIKI-Bench, the first hierarchical benchmark tailored for embodied multi-agent cooperation, featuring three structured levels: agent activation, task planning, and trajectory perception. VIKI-Bench includes diverse robot embodiments, multi-view visual observations, and structured supervision signals to evaluate reasoning grounded in visual inputs. To demonstrate the utility of VIKI-Bench, we propose VIKI-R, a two-stage framework that fine-tunes a pretrained vision-language model (VLM) using Chain-of-Thought annotated demonstrations, followed by reinforcement learning under multi-level reward signals. Our extensive experiments show that VIKI-R significantly outperforms baselines method across all task levels. Furthermore, we show that reinforcement learning enables the emergence of compositional cooperation patterns among heterogeneous agents. Together, VIKI-Bench and VIKI-R offer a unified testbed and method for advancing multi-agent, visual-driven cooperation in embodied AI systems.