Core Insights - Ping An's financial large language model, PingAnGPT-Qwen3-32B, achieved the highest overall score in the CNFinBench evaluation, which benchmarks AI capabilities in the financial sector [1] Group 1: Model Performance - The CNFinBench evaluation included leading models such as GPT-4o and Claude Sonnet 4, as well as Chinese open-source models like DeepSeek-R1 and Qwen3-235B-A22B [1] - PingAnGPT-Qwen3-32B excelled in financial factual reasoning, financial knowledge Q&A, and compliance and risk control, demonstrating high numerical accuracy and logical reasoning [1] Group 2: Business Applications - The model has been deployed across 97 real-world business scenarios, including auto insurance claims, customer service, expense auditing, and intelligent call operations [1] - By enhancing AI capabilities and optimizing model deployment, Ping An aims to provide efficient and high-quality digital financial services [1] Group 3: Company Overview - Ping An is one of the largest financial services companies globally, with over RMB 12 trillion in total assets as of December 2024 [1] - The company ranked 27th in the Forbes Global 2000 list and 47th in the Fortune Global 500 list in 2025, and received an AAA rating in MSCI ESG Ratings [1]
Ping An's Financial LLM Ranks First in CNFinBench Evaluation