Home / Technology / Ant Group Unlocks AI for Finance: Ling-3.0-flash-Fin Open-Sourced
Ant Group Unlocks AI for Finance: Ling-3.0-flash-Fin Open-Sourced
9 Sep
Summary
- Ant Group open-sourced Ling-3.0-flash-Fin, an AI model for financial tasks.
- The model uses a Mixture-of-Experts architecture for efficiency.
- Ant Group also open-sourced FinFIRST, a benchmark for financial search agents.
Ant Group announced the open-sourcing of Ling-3.0-flash-Fin at the 2026 Inclusion·Conference on the Bund. This AI model is tailored for practical financial workflows, emphasizing utility and efficiency in complex tasks. It was developed with input from leading financial institutions to meet stringent compliance and dynamic information handling needs.
The Ling-3.0-flash-Fin model features a Mixture-of-Experts architecture. It has 124 billion total parameters but activates only 5.1 billion per token. This design balances extensive knowledge capacity with low inference costs and high deployment efficiency, performing competitively across numerous financial benchmarks.
Available on OpenRouter and Vercel, with open weights accessible via Hugging Face and ModelScope, Ling-3.0-flash-Fin supports private deployment and integration with search, Python, databases, and spreadsheets. Ant Group also open-sourced FinFIRST, an expert-built benchmark for financial search agents, developed with CICC's investment banking team.