ICPC-Eval: Probing the Frontiers of LLM Reasoning with Competitive Programming Contests

TL;DR

ICPC-Eval is a new benchmark for evaluating large language models' reasoning abilities in competitive programming, featuring realistic problems, robust evaluation tools, and a novel metric to better assess iterative problem-solving skills.

Contribution

The paper introduces ICPC-Eval, a comprehensive benchmark with realistic contest problems, a local evaluation toolkit, and a new metric for assessing reasoning in LLMs, addressing limitations of existing benchmarks.

Findings

Top-tier models rely on multi-turn feedback for reasoning.

Models still lag behind human performance in complex coding tasks.

ICPC-Eval reveals the challenges in evaluating reasoning abilities.

Abstract

With the significant progress of large reasoning models in complex coding and reasoning tasks, existing benchmarks, like LiveCodeBench and CodeElo, are insufficient to evaluate the coding capabilities of large language models (LLMs) in real competition environments. Moreover, current evaluation metrics such as Pass@K fail to capture the reflective abilities of reasoning models. To address these challenges, we propose \textbf{ICPC-Eval}, a top-level competitive coding benchmark designed to probing the frontiers of LLM reasoning. ICPC-Eval includes 118 carefully curated problems from 11 recent ICPC contests held in various regions of the world, offering three key contributions: 1) A challenging realistic ICPC competition scenario, featuring a problem type and difficulty distribution consistent with actual contests. 2) A robust test case generation method and a corresponding local evaluation toolkit, enabling efficient and accurate local evaluation. 3) An effective test-time scaling evaluation metric, Refine@K, which allows iterative repair of solutions based on execution feedback. The results underscore the significant challenge in evaluating complex reasoning abilities: top-tier reasoning models like DeepSeek-R1 often rely on multi-turn code feedback to fully unlock their in-context reasoning potential when compared to non-reasoning counterparts. Furthermore, despite recent advancements in code generation, these models still lag behind top-performing human teams. We release the benchmark at: https://github.com/RUCAIBox/Slow_Thinking_with_LLMs