# FinVector-Market-4B shows large gains on financial tasks after LoRA adaptation

> The model improves exact‑match scores from 14.7% to 40% on FinQA and boosts other task accuracies after a small LoRA fine‑tune.

Oossa · 2026-10-08 · https://oossa.com/en/finvector-market-4b-shows-large-gains-on-financial-tasks-after-lora-adaptation

Researchers led by Alina Khaybullina released FinVector-Market-4B, a version of Qwen3.5‑4B that was adapted with rank‑16 LoRA on 22,000 financial examples. The study compares the base model and the adapted one on a 600‑example benchmark using JSON‑schema prompts. Providing only the schema lifts JSON validity from 0% to 91.3% for the base model. With matched explicit prompting, the adapted model raises FinQA exact‑match accuracy from 14.7% to 40.0% and improves calculator‑expression correctness from 48.0% to 82.7%.

## The facts

- FinVector-Market-4B adapts Qwen/Qwen3.5‑4B with rank‑16 LoRA on 22,000 examples
- Published on 2026-10-08

## Why it matters

The gains suggest small, targeted fine‑tuning can make general‑purpose models much more useful for finance professionals.

## Sources & references

1. [FinVector-Market-4B: A Controlled Study of LoRA Adaptation for Structured Financial Tasks](https://arxiv.org/abs/2610.08882) – arXiv cs.LG, 2026-10-08

Last updated: 2026-10-08