Shanghai University of Finance and Economics (SUFE) unveiled its upgraded "Kuanshi Financial Education Large Model 2.0" alongside a series of achievements in deep integration between artificial intelligence and financial education. The model establishes a continuously updatable "knowledge brain" covering over 2,000 applied economics textbooks, 10,000+ professional books, 10,000+ course Q&A pairs, and 1 million+ high-level academic papers, forming a closed-loop knowledge system.
Developed by SUFE’s newly established AI Finance Development and Service Center, Kuanshi 2.0 serves as the core engine of a four-pronged ecosystem. This includes Fin-R1-Pro, a financial model with professional decision-making capabilities; Fin-Eval, an authoritative evaluation system for financial models; and self-developed AI financial agents tailored for scenario-based applications. Together, they create a full-cycle framework from R&D and capability assessment to industrial deployment, marking SUFE’s advancement in AI-finance integration.
Notably, Kuanshi 2.0 has completed algorithmic registration with China’s Cyberspace Administration, ensuring traceable content generation and verifiable results. It is now deeply embedded in financial teaching, research, and management workflows. Fin-R1-Pro employs a proprietary five-stage data synthesis and filtering strategy, refining over 1 million high-quality financial corpora. Its innovative SRPO algorithm ranks first across 13 authoritative financial datasets and 30 sub-evaluation benchmarks, outperforming general-purpose models like DeepSeek-R1, GPT-4o, and Kimi K2. The Fin-Eval system, now at version 10.0, has expanded to assess financial agents, operational security, high-level task performance, and model attribution, covering six sectors—banking, funds, securities, insurance, trusts, and futures—to align evaluations with real-world practices.
Leveraging Kuanshi 2.0, SUFE has built a universal financial agent system featuring 50+ specialized AI agents. These span conversational learning companions, procedural research assistants, creative tools, collaborative platforms, and self-generating modules, reflecting an evolving capability pathway. By integrating AI with statistics, finance, accounting, and economics, the university is transitioning AI from a mere "auxiliary tool" to a "full-process enabler" in education and research, fostering interdisciplinary innovation and intelligent academic development.