AgentCoMa: A Compositional Benchmark Mixing Commonsense and Mathematical Reasoning in Real-World Scenarios (2026.acl-long)
Copied to clipboard
| Challenge: | brittleness of Large Language Models in reasoningintensive tasks is a problem . current compositional benchmarks focus on *either* commonsense or math reasoning . |
| Approach: | They propose a "Co**mmonsense and Ma**th" benchmark where each compositional task requires a commonsense reasoning step *and* a math reasoning step. |
| Outcome: | The proposed benchmarks show that LLMs can solve both steps in isolation, but their accuracy drops by nearly 30% when the two steps are combined. |
Similar Papers
GroundCocoa: A Benchmark for Evaluating Compositional & Conditional Reasoning in Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing LLMs excel and often surpass human performance on benchmarks, but they are known to falter in simple tasks and under seemingly straightforward circumstances. |
| Approach: | They propose a benchmark to assess compositional and conditional reasoning within a flight booking task. |
| Outcome: | The proposed model outperforms existing models on the flight booking task with a 67% accuracy rate. |
MMAU: A Holistic Benchmark of Agent Capabilities Across Diverse Domains (2025.findings-naacl)
Copied to clipboard
Guoli Yin, Haoping Bai, Shuang Ma, Feng Nan, Yanchao Sun, Zhaoyang Xu, Shen Ma, Jiarui Lu, Xiang Kong, Aonan Zhang, Dian Ang Yap, Yizhe Zhang, Karsten Ahnert, Vik Kamath, Mathias Berglund, Dominic Walsh, Tobias Gindele, Juergen Wiest, Zhengfeng Lai, Xiaoming Simon Wang, Jiulong Shan, Meng Cao, Ruoming Pang, Zirui Wang
| Challenge: | Existing benchmarks focus on specific application scenarios, emphasizing task completion but failing to dissect the underlying skills that drive these outcomes. |
| Approach: | They propose a Massive Multitask Agent Understanding benchmark that evaluates LLMs across five domains and offline tasks. |
| Outcome: | The Massive Multitask Agent Understanding (MMAU) benchmark evaluates models across five domains including Tool-use, Directed Acyclic Graph (DAG) QA, Data Science and Machine Learning coding, Contest-level programming and Mathematics. |
AgentQuest: A Modular Benchmark Framework to Measure Progress and Improve LLM Agents (2024.naacl-demo)
Copied to clipboard
Luca Gioacchini, Giuseppe Siracusano, Davide Sanvito, Kiril Gashteovski, David Friede, Roberto Bifulco, Carolin Lawrence
| Challenge: | Existing benchmarks are narrow and simply compute overall task success. |
| Approach: | They propose a framework where both benchmarks and metrics are modular and easily extensible through well documented and easy-to-use APIs. |
| Outcome: | The proposed framework can track agent progress on two use cases and identify common failure points and refine the agent architecture to obtain a significant performance increase. |
From A and B to A+B: Can Large Language Models Solve Compositional Math Problems? (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies that create problem variants by adding perturbations to a single problem focus on the interaction between problems. |
| Approach: | They propose a pipeline with 98.2% accuracy to combine two original problems with a logical connection and to evaluate LLMs' generalization ability on the compositional problems. |
| Outcome: | The proposed pipeline can combine two original problems with a logical connection to get a new math problem and evaluate its compositional generalization on the compositional problems. |
AgentGym2: Benchmarking Large Language Model Agents in De-Idealized Real-World Environments (2026.acl-long)
Copied to clipboard
Zhiheng Xi, Dingwen Yang, Jiaqi Liu, Jixuan Huang, Honglin Guo, Baodai Huang, Tinggang Chen, Qi Zhang, Zhonghang Lu, Chenyu Liu, Jiajun Sun, Jiazheng Zhang, Dingwei Zhu, Xin Guo, Junzhe Wang, Zhihao Zhang, Yuming Yang, Junjie Ye, Minghe Gao, Dongrui Liu, Jiaming Ji, Guohao Li, Tao Gui, Qi Zhang, Xuanjing Huang
| Challenge: | Existing benchmarks evaluate agents in simplified, idealized settings, relying on pre-packaged tool interfaces, overlooking critical steps, and assume inputs are clean and fully specified. |
| Approach: | They propose a framework that evaluates language agents in simplified, idealized settings . they show that even SOTA systems like Gemini and GPT-5 struggle on AgentGym2 . |
| Outcome: | Experiments on 15 proprietary and open-source models show that even SOTA systems like Gemini and GPT-5 struggle on AgentGym2 . |
Multi-Agent Reasoning Improves Compute Efficiency: Pareto-Optimal Test-Time Scaling (2026.acl-srw)
Copied to clipboard
| Challenge: | Inference methods that prioritize raw performance over cost-effective compute usage are not efficient for real-world applications. |
| Approach: | They evaluate inference scaling strategies to determine their computational efficiency tradeoffs . they find debate and mixture-of-agents outperform self-consistency by 1.3% and 2.7% points . |
| Outcome: | The proposed scaling strategies outperform self-consistency, self-refinement, multi-agent debate and mixture-of-a agents on reasoning tasks. |
MASA: LLM-Driven Multi-Agent Systems for Autoformalization (2025.emnlp-demos)
Copied to clipboard
| Challenge: | This paper presents a framework for building multi-agent systems for autoformalization driven by Large Language Models. |
| Approach: | They propose a framework for building multi-agent systems for autoformalization driven by Large Language Models. |
| Outcome: | The proposed framework leverages collaborative agents to convert natural language statements into formal representations. |
KoCommonGEN v2: A Benchmark for Navigating Korean Commonsense Reasoning Challenges in Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Language models are striving to grasp commonsense reasoning, but they are lacking in Korean commons- ense benchmarks. |
| Approach: | They present a fine-grained benchmark dataset focused on Korean commonsense reasoning that includes multiple-choice questions across seven error categories. |
| Outcome: | The proposed datasets show that LLMs struggle with Korean commonsense reasoning . human accuracy benchmarked at approximately 85%, while GPT-4’s performance lags at about 74%, and other LLM models demonstrate an average accuracy of around 42%. |
Exploring the Compositional Deficiency of Large Language Models in Mathematical Reasoning Through Trap Problems (2024.emnlp-main)
Copied to clipboard
| Challenge: | Current LLMs lack systematic compositionality, and therefore cannot serve as reliable cognitive models. |
| Approach: | They propose to introduce logical traps into the original problems of MATH and GSM8K to investigate the compositionality of large language models in mathematical reasoning. |
| Outcome: | The proposed model can generate infinite combinations from finite learned components. |
Towards Reliable Agents: Benchmarking Customized LLM-Based Retrieval-Augmented Generation Frameworks with Deployment Validation (2025.naacl-industry)
Copied to clipboard
| Challenge: | Existing benchmarks for general-purpose RAG systems, such as CRAG, RGB, MultiHop-RAG, and CRUD-RAGG, are limited and lack a benchmark specifically tailored to evaluate frameworks. |
| Approach: | They evaluated OpenAI’s Assistants API versus a RAG assistant built with Langchain and deployed a system based on benchmark insights as a course assistant over a two-year span. |
| Outcome: | The proposed benchmarks show that domain-specific retrieval impacts response accuracy and highlight key challenges in real-world deployment. |