{"repo":"HiThink-Research/BizFinBench","free":true,"listed":false,"github":"https://github.com/HiThink-Research/BizFinBench","clone":"git clone https://github.com/HiThink-Research/BizFinBench.git","description":"A Business-Driven Real-World Financial Benchmark for Evaluating LLMs","language":"Python","stars":168,"topics":["finance","llm","benchmark","llm-benchmarking","llm-evaluation"],"license":null,"category":"ai-agents","readme_excerpt":"BizFinBench: A Business-Driven Real-World Financial Benchmark for Evaluating LLMs HiThink Research Team Corresponding author, zhangrongjunchen@myhexin.com BizFinBench.v1 (Outdated): 📖 Paper 🏠 Homepage 🤗 Huggingface BizFinBench.v2 (ICML 2026, Yes use this version): 📖 Paper 🏠 Homepage 🤗 Huggingface Large language models excel across general tasks, yet judging their reliability in logic‑heavy, precision‑critical domains such as finance, law and healthcare is still difficult. To address this challenge, we propose BizFinBench , the first benchmark grounded in real-world financial applications. BizFinBench comprises over 100,000+ bilingual (English & Chinese) financial questions, each rooted in real-world business scenarios. 📢 News - 🚀 [01/05/2026] BizFinBench.v2 has been accepted to ICML 2026. - 🚀 [09/01/2026] BizFinBench.v2 is out: 29,578 real-world financial questions so tough that ChatGPT-5 only scores 61.5/100. - 🚀 [04/07/2025] External API support is now live—evaluate BizFinBench with your own endpoints in just a few calls. - 🚀 [16/05/2025] We released BizFinBench.v1 benchmark, the first benchmark grounded in real-world financial applications. 💡 Highlights - 🔥 Benchmark: We propose BizFinBench , the first evaluation benchmark in the financial domain that integrates business-oriented tasks, covering 5 dimensions and 9 categories. It is designed to assess the capacity of LLMs in real-world financial scenarios. - 🔥 Judge model: We design a novel evaluation method, ","default_branch":null,"files":null,"tree":[],"storefront":"/r/HiThink-Research","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/HiThink-Research/BizFinBench/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}