Ant Group Open-Sources Ling-3.0-flash-Fin Financial Model with FinFIRST Benchmark

Ant Group's Ant Ling team open-sourced Ling-3.0-flash-Fin, the first finance-enhanced model in the Ling family, on September 3–4, alongside the FinFIRST benchmark for evaluating financial search agents. Weights are available for download on Hugging Face.

Confirmed

The model has 124B total parameters with only 5.1B activated, uses an efficient MoE architecture with a 256K context window, and was built by continually training Ling 3.0 flash on high-quality financial data
Built on the Bailing V3 architecture, its chat template supports a controllable thinking switch and function calling
The model is designed for real-world financial workflows and can connect to tools like search, Python, databases, and spreadsheets
FinFIRST, open-sourced at the same time, was built by domain experts to evaluate financial search agents

Why it matters

@FellMentKE and @alifcoder both see the most valuable aspect of this release as the combination of "open-source model + public evaluation framework": the open model lowers the barrier to acquiring finance-domain models, while the accompanying benchmark lets developers measure real capability on financial search tasks openly and reproducibly, forming a complete ecosystem loop

2026-09-03 ~ 2026-09-04 · 8 related posts

Primary sources

Ling-3.0-flash-Fin Weights Released: 124B Params, 5.1B Active, 256K Context — Bestlife73 · 2026-09-03 23:44 (UTC+08:00)
Ant Group Open-Sources Finance-Enhanced Model Ling-3.0-flash-Fin with 124B Parameters — niacolhealth · 2026-09-03 23:53 (UTC+08:00)

[source] Ling-3.0-flash-Fin Weights Released: 124B Params, 5.1B Active, 256K Context — Bestlife73 · 2026-09-03
[source] Ant Group Open-Sources Finance-Enhanced Model Ling-3.0-flash-Fin with 124B Parameters — niacolhealth · 2026-09-03
Ling open-sources 124B finance model Ling-3.0-flash-Fin alongside FinFIRST benchmark — alifcoder · 2026-09-04
Ant Group releases FinFIRST, an open benchmark for financial search agents with 12,300 rubric points — FellMentKE · 2026-09-04
FinFIRST details: 123 expert tasks, 701 atomic criteria and 12,300 rubric points for financial search agents — FellMentKE · 2026-09-04
Ant Ling releases Ling-3.0-flash-Fin weights on Hugging Face — FellMentKE · 2026-09-04

2 near-duplicate retellings: FellMentKE · FellMentKE