ModernFinBERT

A fine-tuned financial sentiment analysis model based on ModernBERT, trained on synthetic and real financial data cleaned through an automated AI agentic pipeline. The model covers diverse financial domains including news, tweets, crypto, and macroeconomics, making it the most general-purpose financial sentiment classifier. Benchmark results show superior performance with up to 48% accuracy improvement over existing models across multiple financial datasets.

For private API access or access to even more powerful financial models, contact us at info@tabularis.ai

Quick Start

from transformers import pipeline

# Load model
classifier = pipeline('text-classification', model='tabularisai/ModernFinBERT')

# Test sentences
sentences = [
    "The company reported strong quarterly earnings with revenue growth of 15% year-over-year, exceeding analyst expectations.",
    "Due to rising inflation and supply chain disruptions, the Federal Reserve decided to increase interest rates by 0.75 basis points.",
    "The merger between the two pharmaceutical giants is expected to create significant synergies and reduce operational costs by $2 billion annually."
]

# Evaluate
for i, sentence in enumerate(sentences, 1):
    result = classifier(sentence)
    print(f"Sentence {i}: {result[0]['label']} ({result[0]['score']:.3f})")

Benchmark Results

Dataset	Model	Accuracy	F1-Score	Precision	Recall	ROC-AUC
FIQA	ModernFinBERT	0.80	0.61	0.64	0.88	0.96
FIQA	distilroberta_financial	0.54	0.47	0.61	0.71	0.71
FIQA	finbert	0.48	0.43	0.59	0.66	0.76
FIQA	finbert-tone	0.36	0.36	0.62	0.58	0.77
FIQA	roberta_sentiment	0.36	0.35	0.60	0.58	0.89
Twitter	ModernFinBERT	0.71	0.70	0.68	0.81	0.94
Twitter	distilroberta_financial	0.75	0.71	0.68	0.75	0.87
Twitter	finbert-tone	0.75	0.66	0.68	0.64	0.83
Twitter	finbert	0.73	0.67	0.65	0.70	0.86
Twitter	roberta_sentiment	0.70	0.61	0.63	0.60	0.82
JeanBaptiste	ModernFinBERT	0.74	0.58	0.71	0.56	0.84
JeanBaptiste	distilroberta_financial	0.88	0.79	0.92	0.74	0.86
JeanBaptiste	finbert	0.77	0.68	0.70	0.67	0.88
JeanBaptiste	finbert-tone	0.74	0.60	0.72	0.56	0.86
JeanBaptiste	roberta_sentiment	0.70	0.55	0.79	0.51	0.83

Model Averages Across All Datasets

Model	Accuracy	F1-Score	Precision	Recall	ROC-AUC
ModernFinBERT	0.75	0.63	0.68	0.75	0.91
distilroberta_financial	0.73	0.66	0.73	0.73	0.82
finbert	0.66	0.59	0.65	0.68	0.84
finbert-tone	0.62	0.54	0.68	0.59	0.82
roberta_sentiment	0.59	0.50	0.67	0.56	0.84

Legend:

Bold = Best result per metric per dataset

Italic = Second best result per metric per dataset