Fin_Agent Achieves High Performance in Multilingual Financial NLP Tasks

Fin_Agent utilizes Qwen2.5-32B and BGE-M3 retrieval to tackle financial NER, Japanese implicit communication, and multilingual QA. The system achieves a 0.9521 public F1 for Greek NER and 98.0% accuracy for Japanese implicit communication recognition.