CFA-GPT is a specialized language model fine-tuned on the CFA (Chartered Financial Analyst) curriculum, designed to support finance education and professional development. Built on Meta's Llama 3 8B using QLoRA efficient fine-tuning, it embeds the deep knowledge required for one of the most respected credentials in the global finance industry.
The model was trained on CFA program curriculum materials, financial analysis textbooks, and investment research publications covering the ten core topic areas of the CFA curriculum: ethical and professional standards, quantitative methods, economics, financial reporting and analysis, corporate finance, equity investments, fixed income, derivatives, alternative investments, and portfolio management. This comprehensive training gives CFA-GPT a structured understanding of the principles and methodologies that underpin professional investment analysis.
For finance students, CFA candidates, and professionals seeking to refresh their knowledge, CFA-GPT provides an accessible AI assistant for exploring investment concepts, testing understanding of portfolio theory, and reviewing financial ethics scenarios. While its scope is limited to the CFA curriculum and it is not validated for production financial decision-making, it serves as a powerful educational tool for the finance community.