Reproduce “Quant-Arena: Benchmarking LLM Agents for Quantitative Trading Research” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Houde Qian, Zheng Jiang, Wenyi Li, Chenxi Xu, Xinjie Lin, Yuzhao Peng, Ruoyu Han, Youjie Zheng and 4 more

Quant-Arena tests whether models can turn stock rankings into investable portfolios. It gives models raw records for a 3,042-stock universe, permits five candidate evaluations on 2022–2023 data, then assesses one frozen strategy on sealed 2024 data. All nine models’ highest-scoring iterative configurations trail the market, but GPT-6 Sol at medium effort beats it across all 38 configurations.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.