Financial work depends on trustworthy sources, consistent definitions, accurate calculations and auditable outputs.
Introducing Ling-3.0-flash-Fin, a finance-enhanced version of Ling-3.0-flash, developed with financial institutions and domain experts.
With 124B total and 5.1B active parameters, it supports information retrieval, research, valuation modeling and report preparation across long reports, research materials and complex workbooks.
The model showed competitive results across FinFIRST, FinSearchComp Verified, FinCRAFT, FinanceAgent v1.1/v2, APEX-Agents, SpreadsheetBench v1/v2 and τ³-Banking.
We will open-source the model weights next week.
Financial expertise also needs a strong general foundation.
Ling-3.0-flash-Fin scored 41 on AA Intelligence Index v4.1.1, compared with 38 for Ling-3.0-flash. The result reflects an improvement in broader model capabilities alongside its financial specialization.
Demo 1 | Financial Search
The model reconstructed four first disclosures of Google’s monthly token volume. Across 23 tool calls, it prioritized primary sources, reconciled definitions and dates, and calculated interval increases with a traceable evidence chain.
Demo 2 | Real Financial Research
For Spire, 55 tool calls across 10-K, earnings releases and 8-K filings identified Storage as discontinued operations, rebuilt continuing-operations earnings and validated $17.8M in historical profit—while preserving the forecast boundary.
open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
Why it matters
A domain-tuned open weightsA model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.Full definition → model that runs long tool-call chains over filings and spreadsheets while keeping an auditable source trail, showing that finance specialization did not cost general capability.
Read the full thread on X ↗
