A comprehensive, three-year study conducted by Dr. Alejandro Lopez Lira, an associate professor of finance at the University of Florida, has identified DeepSeek as the top-performing AI model for stock trading. The research aimed to discover which AI models could consistently generate market-beating returns, and after evaluating several popular options, Lopez Lira has pinpointed DeepSeek as the clear leader.
DeepSeek's portfolio has achieved a remarkable 73% return since its launch in 2023, with a year-to-date gain of 35%. This performance outpaces other models like ChatGPT, which, despite a slightly higher cumulative return of 78% since launch, has only achieved 16% year-to-date. Grok's portfolio has returned 54% since inception, with a 19% year-to-date increase. Lopez Lira noted that DeepSeek is also outperforming Claude, though Anthropic's chatbot has been managing a portfolio for a shorter duration.
Lopez Lira's findings suggest that AI's role in trading is rapidly evolving. His experiment, which has been publicly viewable on the Autopilot app, shows DeepSeek's portfolio up 76% against the S&P 500's 27% over the past year. The study also highlights a growing interest in copy-trading AI-managed portfolios, indicating a shift in how retail investors are approaching the market. While AI offers unprecedented speed and data analysis capabilities, experts caution that human oversight remains crucial, especially in navigating unpredictable market shifts.
