跳到正文
arXiv cs.AI· Xianyang Liu, Shangding Gu, Fan Xu, Manxi Wu, Boyi Li, Jun Wang, Costas Spanos, Dawn Song·· 4 小时前AI 评分45

AgenticPay:面向 LLM 多智能体商务谈判的多模态基准

From Chats to Markets: AgenticPay for LLM-Powered Negotiation in Multi-Agent Commerce

AI 导读

研究者发布多智能体商务谈判基准 AgenticPay,涵盖电商、外卖、网约车、租房四类场景共 160 个多模态任务和八种市场拓扑,评测模型通过多轮对话协商价格、租期等条款。

正文 · 原文

View PDF HTML (experimental)

Abstract:Large language model agents are increasingly expected to negotiate and transact autonomously, yet existing benchmarks rarely combine multimodal product grounding, multi-issue bargaining, and competitive multi-agent markets. We introduce AgenticPay, a unified benchmark for evaluating multimodal agents' ability to reach high-welfare, multidimensional agreements in realistic markets. AgenticPay includes 160 multimodal tasks across four business scenarios (E-commerce, Food Delivery, Ride-hailing, and Apartment Rental) and eight market topologies, from bilateral bargaining to many-to-many competition. Agents interpret product images, infer opponents' hidden preferences, and negotiate binding contract terms such as price, lease duration, return policy, and delivery speed through multi-round natural-language dialogue. A utility-based framework evaluates agreement quality and joint welfare. Experiments with proprietary (GPT-5.4, Claude Sonnet 4.6, Gemini 3.1 Pro Preview) and open-weight (Qwen3-VL-32B-Instruct, InternVL3-38B) multimodal models reveal substantial limitations in value creation and market reasoning. Even the strongest model, Gemini 3.1 Pro Preview, achieves a GlobalScore of only 42.3/100. As markets shift from bilateral to multi-sided, average GlobalScore falls by 5.7 points, with open-weight models declining by up to 11.5 points. AgenticPay provides a testbed for evaluating multimodal agents in realistic commerce. Code and data: this https URL.
Comments: Accepted at NeurIPS 2026 (Main Track)
Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2602.06008 [cs.AI]
  (or arXiv:2602.06008v2 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2602.06008

arXiv-issued DOI via DataCite

Submission history

From: Xianyang Liu [view email]
[v1] Thu, 5 Feb 2026 18:50:36 UTC (1,131 KB)
[v2] Fri, 13 Nov 2026 06:36:34 UTC (2,056 KB)

来源:arXiv cs.AI · arxiv.org