Overview

Fin-o1-8B is an 8 billion parameter language model developed by TheFinAI, built upon the Qwen3-8B architecture. It has been specifically fine-tuned using Supervised Fine-Tuning (SFT) and Grouped Reinforcement Learning with Policy Optimization (GRPO) to excel in financial reasoning tasks. The training utilized a specialized dataset, TheFinAI/FinCoT, which is derived from various financial benchmarks including FinQA, TATQA, DocMath-Eval, Econ-Logic, and BizBench-QA.

Key Capabilities

• Enhanced Financial Reasoning: Optimized for complex financial mathematical reasoning and analytical tasks.
• Specialized Training: Benefits from SFT and GRPO on a diverse financial dataset, improving its understanding and generation of financial insights.
• Qwen3-8B Base: Inherits the robust capabilities and tokenizer of its Qwen3-8B base model.

Good For

• Financial Analysis: Applications requiring detailed financial calculations and reasoning.
• Quantitative Finance: Tasks involving the interpretation and processing of financial data.
• Research: Academic or industry research focused on financial language understanding and generation.