Finatune
← Fine-Tuning for Finance
Fine-Tuned Finance Models

Fine-Tuned Financial AI Models

Open-source and community fine-tuned models for financial services — sentiment analysis, document analysis, embeddings, and financial reasoning, all available on Hugging Face.

12 Models6 CategoriesFree & Open Source
12 models
Sentiment AnalysisActive

FinBERT

ProsusAI / Hugging Face

The most widely used financial NLP model — fine-tuned BERT for financial sentiment analysis with 5M+ monthly downloads on Hugging Face.

BERT-largeApache-2.05M+ monthly
Full fine-tuning on BERT
Sentiment AnalysisResearch

Bloomberg FinBERT

Bloomberg

Bloomberg's fine-tuned BERT model trained on proprietary financial news corpus — superior financial phrase classification for Bloomberg Terminal sentiment signals.

BERT-baseCC-BY-SA-3.0500K+ monthly
Fine-tuning on Bloomberg financial corpus
General FinanceResearch

FinGPT v3.3

AI4Finance Foundation

Open-source financial LLM from AI4Finance Foundation — LoRA fine-tuned Llama for financial NLP tasks with 14K GitHub stars and MIT license.

Llama 3 / Llama 2MIT100K+ monthly
LoRA fine-tuning on Llama
Document AnalysisActive

SEC-BERT

Community (Loukas et al.)

BERT fine-tuned on 260K SEC filings — the best open-source model for SEC EDGAR document analysis, risk factor extraction, and regulatory disclosure classification.

BERT-baseApache-2.0200K+ monthly
Pre-training + fine-tuning on SEC filings
EmbeddingsActive

Finance-Embedding (Fin-E5)

Community

Open-source embedding model fine-tuned on financial document pairs — the best free alternative to Voyage Finance-2 for financial RAG pipelines and document search.

E5-LargeMIT1M+ monthly
Contrastive fine-tuning on financial pairs
General FinanceActive

FinMistral

Community

QLoRA fine-tuned Mistral 7B for financial tasks — Apache 2.0 license, runs on a single GPU, ideal for EU financial institutions wanting GDPR-compliant local finance AI.

Mistral 7BApache-2.050K+ monthly
QLoRA on Mistral 7B
Compliance & AMLResearch

AML-BERT

Community / Research

RoBERTa fine-tuned for AML compliance tasks — classifies financial transactions, supports SAR narrative generation, and detects suspicious financial crime patterns.

RoBERTa-baseApache-2.020K+ monthly
Fine-tuning RoBERTa on AML datasets
Document AnalysisActive

TaxBERT

Community (Chalkidis et al.)

Legal-BERT fine-tuned on tax documents and EU directives — the best open-source model for tax compliance document classification and European regulatory text analysis.

RoBERTa-largeApache-2.0300K+ monthly
Fine-tuning on legal and tax documents
Document AnalysisActive

EarningsCall-LLM

Community

LoRA fine-tuned Llama for earnings call analysis — extracts forward guidance, executive sentiment, and key financial metrics from earnings transcripts with higher accuracy than general-purpose LLMs.

Llama 3.1 8BMIT15K+ monthly
LoRA on Llama 3.1 8B
Financial ReasoningResearch

Finance-LLM

Community (Zhao et al.)

Instruction fine-tuned Llama 2 on 150K financial Q&A pairs — strong financial reasoning and document understanding for research and prototyping financial AI applications.

Llama 2 13BMIT30K+ monthly
Instruction fine-tuning on Llama 2
Financial ReasoningResearch

CFA-GPT

Community

QLoRA fine-tuned Llama 3 on CFA curriculum — investment analysis, portfolio theory, and financial ethics knowledge for finance education and professional development.

Llama 3 8BMIT10K+ monthly
QLoRA on Llama 3 8B
EmbeddingsActive

FinLang Embedding

Community (FinLang AI)

Sentence-BERT fine-tuned on Investopedia and financial glossaries — best open-source embedding model for financial education RAG pipelines and finance concept semantic search.

Sentence-BERTMIT100K+ monthly
Contrastive fine-tuning on finance pairs

Want to Build Your Own Fine-Tuned Finance Model?

Explore our step-by-step fine-tuning guides covering LoRA, QLoRA, dataset preparation, and compliance frameworks for regulated financial institutions.

Browse Fine-Tuning Guides →