4 LLMs Tested in Codex, Claude Code, Hermes & OpenClaw (FinAI)
Skills:
LLM Engineering90%
Should you combined Claude CODE with GPT-5.4 or is Codex with Sonnet 4.6 better? Which LLMs performs best (like a Qwen 27B or a 397B) with open source agent frameworks like OpenClaw or Hermes?
After 32000 hours of NVIDIA GPU we have performance data on all 20 combinations! And the winner is ....
By the way: Thank you to NVIDIA to provide 32000 hours of testing for the researcher to experiment and validate the performance of FinTech and FinAI on real world financial data and the current performance of the best AI systems for financial market predictions (maximum risks).
all rights w/ authors:
"HERCULEAN: An Agentic Benchmark for Financial Intelligence"
Xueqing Peng1, Zhuohan Xie9, Yupeng Cao4, Haohang Li4, Lingfei Qian1, Yan Wang1, Vincent Jim Zhang1,
Huan He1, Xuguang Ai1, Linhai Ma1, Ruoyu Xiang6, Yueru He3, Yi Han7, Shuyao Wang1, Yuqing Guo1,
Mingyang Jiang1, Yilun Zhao2, Youzhong Dong1, Xiaoyu Wang6, Yankai Chen9,20, Ye Yuan20,21,
Qiyuan Zhang9, Fuyuan Lyu20,21, Haolun Wu20,21, Yonghan Yang9, Zichen Zhao9, Yuyang Dai1,
Fan Zhang9, Rania Elbadry9, Ayesha Gull1, Muhammad Usman Safder1, Nuo Chen16, Fengbin Zhu16,
Tianshi Cai14, Zimu Wang14, Polydoros Giannouris18, Yuechen Jiang18, Zhiwei Liu18, Mohsinul Kabir18,
Yuyan Wang18, Yixiang Zheng18, Yangyang Yu4, Weijin Liu4, Wenbo Cao1, Anke Xu1, Peng Lu10,
Jerry Huang10, Mingquan Lin11, Prayag Tiwari17, Yijia Zhao12, Victor Gutierrez Basulto19,
Xiao-Yang Liu3, Kaleb E. Smith5, Jiahuan Pei15, Arman Cohan2, Jimin Huang1,10, Yuehua Tang8,
Alejandro Lopez-Lira8, Xi Chen6, Xue Liu9,20,21, Junichi Tsujii13, Jian-Yun Nie10, Sophia Ananiadou18
1The Fin AI, 2Yale University, 3Columbia University, 4Stevens Institute of Technology, 5NVIDIA,
6New York University, 7Georgia Institute of Technology, 8University of Florida, 9MBZUAI,
10Université de Montréal, 11University of Minnesota, 12University of Massachusetts Boston,
13National Institute of Advanced Industrial Science and Technology, 14University of Liverpool,
15Vrije Univers
Watch on YouTube ↗
(saves to browser)
Sign in to unlock AI tutor explanation · ⚡30
More on: LLM Engineering
View skill →Related AI Lessons
⚡
⚡
⚡
⚡
Tokenyst Review: Track Claude Code API Costs Before the Bill Lands
Dev.to AI
Cursor vs GitHub Copilot: Which AI Coding Assistant Ships Faster in 2026?
Dev.to · pickuma
Forget Watching 100 AI Tutorials. Do This Instead
Medium · AI
Spec-Driven Development: How AI Is Flipping the Script on Software Engineering
Medium · AI
🎓
Tutor Explanation
DeepCamp AI