Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.20502 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912979553878016 |
|---|---|
| author | Wang, Xiao Wang, Jia Wang, Yijie Dang, Pengtao Cao, Sha Zhang, Chi |
| author_facet | Wang, Xiao Wang, Jia Wang, Yijie Dang, Pengtao Cao, Sha Zhang, Chi |
| contents | Large language models (LLMs) have achieved impressive results in natural language understanding, yet their reasoning capabilities remain limited when operating as single agents. Multi-Agent Debate (MAD) has been proposed to address this limitation by enabling collaborative reasoning among multiple models in a round-table debate manner. While effective, MAD introduces substantial computational overhead due to the number of agents involved and the frequent communication required. In this paper, we propose MARS (Multi-Agent Review System), a role-based collaboration framework inspired by the review process. In MARS, an author agent generates an initial solution, reviewer agents provide decisions and comments independently, and a meta-reviewer integrates the feedback to make the final decision and guide further revision. This design enhances reasoning quality while avoiding costly reviewer-to-reviewer interactions, thereby controlling token consumption and inference time. We compared MARS with both MAD and other state-of-the-art reasoning strategies across multiple benchmarks. Extensive experiments with different LLMs show that MARS matches the accuracy of MAD while reducing both token usage and inference time by approximately 50\%. Code is available at https://github.com/xwang97/MARS. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_20502 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MARS: toward more efficient multi-agent collaboration for LLM reasoning Wang, Xiao Wang, Jia Wang, Yijie Dang, Pengtao Cao, Sha Zhang, Chi Computation and Language Artificial Intelligence Large language models (LLMs) have achieved impressive results in natural language understanding, yet their reasoning capabilities remain limited when operating as single agents. Multi-Agent Debate (MAD) has been proposed to address this limitation by enabling collaborative reasoning among multiple models in a round-table debate manner. While effective, MAD introduces substantial computational overhead due to the number of agents involved and the frequent communication required. In this paper, we propose MARS (Multi-Agent Review System), a role-based collaboration framework inspired by the review process. In MARS, an author agent generates an initial solution, reviewer agents provide decisions and comments independently, and a meta-reviewer integrates the feedback to make the final decision and guide further revision. This design enhances reasoning quality while avoiding costly reviewer-to-reviewer interactions, thereby controlling token consumption and inference time. We compared MARS with both MAD and other state-of-the-art reasoning strategies across multiple benchmarks. Extensive experiments with different LLMs show that MARS matches the accuracy of MAD while reducing both token usage and inference time by approximately 50\%. Code is available at https://github.com/xwang97/MARS. |
| title | MARS: toward more efficient multi-agent collaboration for LLM reasoning |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2509.20502 |