Ant's Ling-3.0-flash-Fin ships with FinFIRST: open model plus open benchmark in one release
FellMentKE · x · 2026-09-04
Ant Group's Ling team released the open Ling-3.0-flash-Fin model alongside FinFIRST, an open benchmark for financial search agents. The model is built for financial workflows and can connect to search, Python, databases, and spreadsheets. The author's key point: releasing both the model and its evaluation benchmark in public is a stronger combination than open-sourcing a model alone, giving teams a complete setup to test their own workflow ideas end to end.
More from Models
- Astra Priced at $10/M Input and $50/M Output Tokens — saln1 · 2026-09-04
- GPT-6 Astra vs Gemini 3.8 Flash: user posts head-to-head comparison — Able-Line2683 · 2026-09-04
- GPT-6 Astra pricing leaked: $10/M input, $50/M output tokens, near-perfect ARC-AGI-3 — thesaraharminta · 2026-09-04
- GPT-6 Astra hits 97.4% on FrontierMath Tier 4, trained on 100k+ GPUs — ChrisGPT · 2026-09-04
- Blogger speculates a $50 GPT-6 Astra subscription tier is possible — scaling01 · 2026-09-04
- Gemini 3.8 Flash flops on warden bench: worst cost-per-finding of tested models — zeeg · 2026-09-04