dorsal/arxiv
View SchemaPriceSeer: Evaluating Large Language Models in Real-Time Stock Prediction
| Authors | Bohan Liang, Zijian Chen, Qi Jia, Kaiwei Zhang, Kaiyuan Ji, Guangtao Zhai |
|---|---|
| Categories | |
| ArXiv ID | 2601.06088vv1 |
| URL | https://arxiv.org/abs/2601.06088 |
| License | http://creativecommons.org/licenses/by/4.0/ |
Abstract
Stock prediction, a subject closely related to people's investment activities in fully dynamic and live environments, has been widely studied. Current large language models (LLMs) have shown remarkable potential in various domains, exhibiting expert-level performance through advanced reasoning and contextual understanding. In this paper, we introduce PriceSeer, a live, dynamic, and data-uncontaminated benchmark specifically designed for LLMs performing stock prediction tasks. Specifically, PriceSeer includes 110 U.S. stocks from 11 industrial sectors, with each containing 249 historical data points. Our benchmark implements both internal and external information expansion, where LLMs receive extra financial indicators, news, and fake news to perform stock price prediction. We evaluate six cutting-edge LLMs under different prediction horizons, demonstrating their potential in generating investment strategies after obtaining accurate price predictions for different sectors. Additionally, we provide analyses of LLMs' suboptimal performance in long-term predictions, including the vulnerability to fake news and specific industries. The code and evaluation data will be open-sourced at https://github.com/BobLiang2113/PriceSeer.
{
"annotation_id": "7f7c0b56-b360-4182-aefa-457885386050",
"date_created": "2026-02-17T05:53:04.894000Z",
"date_modified": "2026-02-17T05:53:04.894000Z",
"file_hash": "10c81e1d43f9e71a14974f41a90eb91cac092a85ade3f126657a001a85787bb8",
"private": false,
"record": {
"abstract": "Stock prediction, a subject closely related to people\u0027s investment activities in fully dynamic and live environments, has been widely studied. Current large language models (LLMs) have shown remarkable potential in various domains, exhibiting expert-level performance through advanced reasoning and contextual understanding. In this paper, we introduce PriceSeer, a live, dynamic, and data-uncontaminated benchmark specifically designed for LLMs performing stock prediction tasks. Specifically, PriceSeer includes 110 U.S. stocks from 11 industrial sectors, with each containing 249 historical data points. Our benchmark implements both internal and external information expansion, where LLMs receive extra financial indicators, news, and fake news to perform stock price prediction. We evaluate six cutting-edge LLMs under different prediction horizons, demonstrating their potential in generating investment strategies after obtaining accurate price predictions for different sectors. Additionally, we provide analyses of LLMs\u0027 suboptimal performance in long-term predictions, including the vulnerability to fake news and specific industries. The code and evaluation data will be open-sourced at https://github.com/BobLiang2113/PriceSeer.",
"arxiv_id": "2601.06088",
"authors": [
"Bohan Liang",
"Zijian Chen",
"Qi Jia",
"Kaiwei Zhang",
"Kaiyuan Ji",
"Guangtao Zhai"
],
"categories": [
"q-fin.ST",
"cs.LG"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"title": "PriceSeer: Evaluating Large Language Models in Real-Time Stock Prediction",
"url": "https://arxiv.org/abs/2601.06088",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "903a31ce-e4e2-4b4d-982d-45356ec23ced",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}