Agent-RewardBench: Towards a Unified Benchmark for Reward Modeling across Perception, Planning, and Safety in Real-World Multimodal Agents

• URL: http://arxiv.org/abs/2506.21252v1
• Date: Thu, 26 Jun 2025 13:36:12 GMT
• Title: Agent-RewardBench: Towards a Unified Benchmark for Reward Modeling across Perception, Planning, and Safety in Real-World Multimodal Agents
• Authors: Tianyi Men, Zhuoran Jin, Pengfei Cao, Yubo Chen, Kang Liu, Jun Zhao,
• Abstract summary: multimodal agents show promise in real-world tasks like web navigation and embodied intelligence.<n>Due to limitations in a lack of external feedback, these agents struggle with self-correction and generalization.<n>There is no clear on how to select reward models for agents.
• Score: 19.015202590038996
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: As Multimodal Large Language Models (MLLMs) advance, multimodal agents show promise in real-world tasks like web navigation and embodied intelligence. However, due to limitations in a lack of external feedback, these agents struggle with self-correction and generalization. A promising approach is to use reward models as external feedback, but there is no clear on how to select reward models for agents. Thus, there is an urgent need to build a reward bench targeted at agents. To address these challenges, we propose Agent-RewardBench, a benchmark designed to evaluate reward modeling ability in MLLMs. The benchmark is characterized by three key features: (1) Multiple dimensions and real-world agent scenarios evaluation. It covers perception, planning, and safety with 7 scenarios; (2) Step-level reward evaluation. It allows for the assessment of agent capabilities at the individual steps of a task, providing a more granular view of performance during the planning process; and (3) Appropriately difficulty and high-quality. We carefully sample from 10 diverse models, difficulty control to maintain task challenges, and manual verification to ensure the integrity of the data. Experiments demonstrate that even state-of-the-art multimodal models show limited performance, highlighting the need for specialized training in agent reward modeling. Code is available at github.

Related papers

• Agent-X: Evaluating Deep Multimodal Reasoning in Vision-Centric Agentic Tasks [94.19506319646376]
  We introduce Agent-X, a benchmark for evaluating vision-centric agents in real-world, multimodal settings.<n>Agent-X features 828 agentic tasks with authentic visual contexts, including images, multi-image comparisons, videos, and instructional text.<n>Our results reveal that even the best-performing models, including GPT, Gemini, and Qwen families, struggle to solve multi-step vision tasks.
  arXiv  Detail & Related papers  (2025-05-30T17:59:53Z)
• Boosting Virtual Agent Learning and Reasoning: A Step-Wise, Multi-Dimensional, and Generalist Reward Model with Benchmark [72.46357004059661]
  Generalist Virtual Agents (GVAs) have shown significant promise in autonomous task execution.<n>To address these challenges, we propose Similar, a Step-Wise Multi-Dimensional Generalist Reward Model.<n>Similar offers fine-grained signals for agent training and can choose better action for inference-time scaling.
  arXiv  Detail & Related papers  (2025-03-24T13:30:47Z)
• VisualPRM: An Effective Process Reward Model for Multimodal Reasoning [76.35753243272521]
  We introduce VisualPRM, which improves the reasoning abilities of existing Multimodal Large Language Models (MLLMs)<n>Our model achieves a 5.9-point improvement across seven multimodal reasoning benchmarks.<n>For the evaluation of multimodal PRMs, we propose VisualProcessBench, a benchmark with human-annotated step-wise correctness labels.
  arXiv  Detail & Related papers  (2025-03-13T12:03:37Z)
• Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systems [54.4392552373835]
  Reward models (RMs) are crucial for the training and inference-time scaling up of large language models (LLMs)<n>We propose agentic reward modeling, a reward system that combines reward models with verifiable correctness signals to provide reliable rewards.<n>We conduct comprehensive experiments on existing reward model benchmarks and inference time best-of-n searches on real-world downstream tasks.
  arXiv  Detail & Related papers  (2025-02-26T17:19:12Z)
• AgentRM: Enhancing Agent Generalization with Reward Modeling [78.52623118224385]
  We find that finetuning a reward model to guide the policy model is more robust than directly finetuning the policy model.<n>We propose AgentRM, a generalizable reward model, to guide the policy model for effective test-time search.
  arXiv  Detail & Related papers  (2025-02-25T17:58:02Z)
• Multimodal RewardBench: Holistic Evaluation of Reward Models for Vision Language Models [82.92771279118888]
  We introduce Multimodal RewardBench, an expert-annotated benchmark for evaluating multimodal reward models.<n>Our dataset comprises 5,211 annotated (prompt, chosen response, rejected response) triplets collected from various vision-language models.<n>We find that even the top-performing models, Gemini 1.5 Pro and Claude 3.5 Sonnet, achieve only 72% overall accuracy.
  arXiv  Detail & Related papers  (2025-02-20T01:48:13Z)
• Towards Robust Multi-Modal Reasoning via Model Selection [7.6621866737827045]
  LLM serves as the "brain" of the agent, orchestrating multiple tools for collaborative multi-step task solving. We propose the $textitM3$ framework as a plug-in with negligible runtime overhead at test-time. Our experiments reveal that our framework enables dynamic model selection, considering both user inputs and subtask dependencies.
  arXiv  Detail & Related papers  (2023-10-12T16:06:18Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.