Can Language Models Solve Olympiad Programming?

TL;DR

This paper introduces the USACO benchmark to evaluate language models on Olympiad programming problems, revealing current limitations and potential improvements through targeted hints and advanced inference methods.

Contribution

It provides the first comprehensive benchmark for LM performance on Olympiad problems, along with new inference techniques and insights from human-in-the-loop studies.

Findings

GPT-4 achieves 8.7% pass@1 accuracy with zero-shot prompting.

Best inference method improves accuracy to 20.2%.

Targeted hints enable solving most previously unsolvable problems.

Abstract

Computing olympiads contain some of the most challenging problems for humans, requiring complex algorithmic reasoning, puzzle solving, in addition to generating efficient code. However, it has been understudied as a domain to evaluate language models (LMs). In this paper, we introduce the USACO benchmark with 307 problems from the USA Computing Olympiad, along with high-quality unit tests, reference code, and official analyses for each problem. These resources enable us to construct and test a range of LM inference methods for competitive programming for the first time. We find GPT-4 only achieves a 8.7% pass@1 accuracy with zero-shot chain-of-thought prompting, and our best inference method improves it to 20.2% using a combination of self-reflection and retrieval over episodic knowledge. However, this is far from solving the benchmark. To better understand the remaining challenges, we design a novel human-in-the-loop study and surprisingly find that a small number of targeted hints enable GPT-4 to solve 13 out of 15 problems previously unsolvable by any model and method. Our benchmark, baseline methods, quantitative results, and qualitative analysis serve as an initial step toward LMs with grounded, creative, and algorithmic reasoning.