SciCode: A Research Coding Benchmark Curated by Scientists

TL;DR

SciCode is a curated benchmark of 338 scientific research coding problems from 16 fields, designed to evaluate language models' ability to solve complex, real-world scientific tasks, revealing current limitations.

Contribution

This paper introduces SciCode, a novel, scientist-curated benchmark for evaluating AI models on real scientific coding problems across multiple disciplines.

Findings

Claude3.5-Sonnet solves only 4.6% of problems in realistic settings

SciCode highlights current AI limitations in scientific problem-solving

Benchmark facilitates future development of scientific AI tools

Abstract

Since language models (LMs) now outperform average humans on many challenging tasks, it has become increasingly difficult to develop challenging, high-quality, and realistic evaluations. We address this issue by examining LMs' capabilities to generate code for solving real scientific research problems. Incorporating input from scientists and AI researchers in 16 diverse natural science sub-fields, including mathematics, physics, chemistry, biology, and materials science, we created a scientist-curated coding benchmark, SciCode. The problems in SciCode naturally factorize into multiple subproblems, each involving knowledge recall, reasoning, and code synthesis. In total, SciCode contains 338 subproblems decomposed from 80 challenging main problems. It offers optional descriptions specifying useful scientific background information and scientist-annotated gold-standard solutions and test cases for evaluation. Claude3.5-Sonnet, the best-performing model among those tested, can solve only 4.6% of the problems in the most realistic setting. We believe that SciCode demonstrates both contemporary LMs' progress towards becoming helpful scientific assistants and sheds light on the development and evaluation of scientific AI in the future.