CCTU: A Benchmark for Tool Use under Complex Constraints

TL;DR

CCTU is a comprehensive benchmark designed to evaluate large language models' ability to use tools under complex, multi-faceted constraints, revealing significant limitations in current models' adherence and self-refinement capabilities.

Contribution

We introduce CCTU, a novel benchmark with detailed constraint categories and validation tools, enabling systematic evaluation of LLM tool use under complex constraints.

Findings

Models rarely achieve high success rates under strict constraints.

Over 50% of constraint violations occur, especially in resource and response categories.

Models show limited self-refinement even with detailed feedback.

Abstract

Solving problems through tool use under explicit constraints constitutes a highly challenging yet unavoidable scenario for large language models (LLMs), requiring capabilities such as function calling, instruction following, and self-refinement. However, progress has been hindered by the absence of dedicated evaluations. To address this, we introduce CCTU, a benchmark for evaluating LLM tool use under complex constraints. CCTU is grounded in a taxonomy of 12 constraint categories spanning four dimensions (i.e., resource, behavior, toolset, and response). The benchmark comprises 200 carefully curated and challenging test cases across diverse tool-use scenarios, each involving an average of seven constraint types and an average prompt length exceeding 4,700 tokens. To enable reliable evaluation, we develop an executable constraint validation module that performs step-level validation and enforces compliance during multi-turn interactions between models and their environments. We evaluate nine state-of-the-art LLMs in both thinking and non-thinking modes. Results indicate that when strict adherence to all constraints is required, no model achieves a task completion rate above 20%. Further analysis reveals that models violate constraints in over 50% of cases, particularly in the resource and response dimensions. Moreover, LLMs demonstrate limited capacity for self-refinement even after receiving detailed feedback on constraint violations, highlighting a critical bottleneck in the development of robust tool-use agents. To facilitate future research, we release the data and code.