Open-source financial large language models for sentiment analysis and market forecasting.
Adaptable financial LLMs for sentiment analysis, robo-advising, and market forecasting.
FinGPT provides open-source financial large language models designed to bypass the high costs of proprietary models like BloombergGPT. It suits quantitative researchers and developers who need lightweight adaptation and RLHF capabilities for financial tasks.
The toolkit delivers instruction-tuned models and data pipelines specifically for financial sentiment analysis, multi-task LLMs, and automated stock price forecasting. It supports lightweight fine-tuning using methods like LoRA to incorporate fresh market data rapidly.
It suits data scientists and finance professionals with machine learning skills who want to customize financial LLMs locally or via cloud APIs. It does not suit non-technical users seeking a ready-to-use commercial trading platform without coding.
Adapt models swiftly using methods like LoRA to incorporate new financial data at a fraction of full retraining costs.
Predict next-week stock price movements using ticker symbols, past market news, and basic financials.
Run state-of-the-art financial sentiment analysis models trained on news and tweets datasets.
Incorporate reinforcement learning from human feedback to learn individual risk-aversion levels and investing habits.
Free under its MIT license, with no paid hosted option provided by the project. The only cost is running it yourself via cloud APIs or your own GPU hardware.
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
BloombergGPT is proprietary and costly to retrain from scratch, whereas FinGPT is open-source, relies on lightweight adaptation for lower costs, and includes RLHF.
Running local models requires a GPU, such as an RTX 3090 for sentiment analysis models, or you can use cloud APIs.
FinGPT-Forecaster is trained on US DOW 30 data; for other markets, you must collect target data and fine-tune the model.
See Zekai first in Google