Where Larger Models Excel: The Primacy of Constraint-Guided Reasoning
For researchers and practitioners building or selecting LLMs for reasoning tasks, this work provides a concrete, data-driven taxonomy of why larger models perform better, revealing that the advantage is not uniform but centered on constraint handling.
Larger language models consistently outperform smaller ones on reasoning benchmarks (e.g., Qwen3-32B beats Qwen3-8B by 6.43%; GPT-OSS-120B beats GPT-OSS-20B by 7.38%). The paper identifies a systematic taxonomy of reasoning advantages, with a recurring theme that larger models excel at constraint-guided reasoning—identifying and organizing constraints to rule out infeasible paths and verify steps.
Larger language models consistently outperform smaller ones on reasoning benchmarks, yet the reasoning differences underlying this gap remain underexplored. Across benchmarks in mathematics, physics, chemistry, and programming, we observe stable performance gaps: averaged over datasets, Qwen3-32B outperforms Qwen3-8B by 6.43%, while GPT-OSS-120B exceeds GPT-OSS-20B by 7.38%. To study the reasoning differences behind these gains, we develop AdvCluster, an automated framework that identifies questions where the larger model shows a stable advantage, extracts fine-grained advantage descriptions from paired reasoning traces produced by larger and smaller models, and organizes them through semantic clustering with quantitative evaluation and selection guided by a reviewer model. Our analysis yields a systematic taxonomy of larger model reasoning advantages, spanning both common advantages that recur across domains and specialized advantages associated with particular domains. Across these patterns, a recurring theme is Constraint-Guided Reasoning: larger models are better at identifying explicit and implicit constraints, organizing them into structured reasoning, and using them to rule out infeasible paths and verify intermediate steps.