Numerai

Hedge Fund Quality Data

Numerai provides clean, regularized, and obfuscated data that is free to use. Each row includes an id (stock at a specific time era), features describing quantitative attributes, and a target representing abstract future performance.

Access ready-to-use data without licensing fees, enabling immediate model development.

Machine Learning for Stock Prediction

Build models using example Python and R scripts. The provided example uses XGBoost regression to train on features and predict the target, then outputs predictions for submission.

Quickly prototype and train predictive models with standard ML tools.

Submit and Stake Models

Submit your model's predictions to the tournament. Stake your best models with NMR cryptocurrency to earn or burn crypto based on performance, building reputation on the leaderboard.

Earn cryptocurrency rewards and climb the leaderboard by staking successful models.

Join a Network of Data Scientists

Numerai combines thousands of models into a meta model. The platform supports a community with forums, Discord, GitHub, and leaderboards, and has paid out $19.6M to data scientists.

Collaborate and compete with a global community while earning from your contributions.

AI Scientist Tools

Numerai offers API keys and MCP setup for AI assistants like Codex, Claude, and Cursor. Clone example-scripts and use the install script to prompt Codex to find the best neural network architecture.

Leverage frontier AI to automate and optimize model discovery.

Backed by Notable Investors

Numerai runs a quantitative global equity market neutral hedge fund, backed by Union Square Ventures and the co-founder of Renaissance Technologies. It is designed for institutional investors.

Participate in a professionally backed fund with a unique crowdsourced approach.

• Hedge fund quality dataNumerai provides clean, regularized, and obfuscated data that is designed to be usable right away and can be given out for free. Each id corresponds to a stock at a specific time era, with features describing quantitative attributes and a target representing an abstract measure of future performance.
• Example scripts and toolsClone the example-scripts repository to get everything you need to start. Example Python and R scripts are included, along with an MCP installer for Codex, Claude, or Cursor. You can also use the provided command to prompt Codex to find the best neural network architecture to predict the target.
• Staking and cryptocurrency rewardsData scientists can stake their best models with NMR, Numerai's cryptocurrency, to earn or burn cryptocurrency based on model performance. This creates a reputation system and a leaderboard for participants.
• Backed by notable investorsNumerai is backed by Union Square Ventures and the co-founder of Renaissance Technologies. It runs a quantitative global equity market neutral hedge fund designed for institutional investors.

• Not for most investorsNumerai's hedge fund is unsuitable for most investors. It is designed for institutional investors, though some high net worth accredited individuals may qualify.
• Data is obfuscatedThe data provided is obfuscated, meaning feature names and target values are abstracted. This is intentional to allow free distribution, but it means you cannot directly interpret the financial meaning of the data.
• Model performance is not guaranteedWhile the example script shows a simple XGBoost model, achieving Grandmaster status or earning cryptocurrency is not guaranteed. Performance depends on your model's predictive ability and the tournament rules.
• Contact and careersFor general enquiries, contact contact@numer.ai. For investor inquiries, use investing@numer.ai. Careers are listed on Wellfound. There is no mention of any certifications or integrations beyond the example scripts and MCP installers.

Predict Stock Market Returns

Use frontier AI to predict the stock market by building models on Numerai's clean, regularized, and obfuscated hedge fund quality data. Each row contains features describing quantitative attributes of a stock at a specific time era, and the target represents an abstract measure of future performance.

Train Machine Learning Models

Apply machine learning to predict the stock market using example Python and R scripts. The provided example uses XGBoost regression to train on features and target data, then outputs predictions for tournament data. Clone the example-scripts repository to get everything you need to start.

Stake Models to Earn Cryptocurrency

Submit your models to predict the stock market and build reputation on the leaderboard. Stake your best models with NMR cryptocurrency to earn or burn crypto based on performance. Over $19.6 million has been paid to data scientists so far.

Contribute to a Meta Model Hedge Fund

Join the network and help build the world's last hedge fund. Numerai combines thousands of models from data scientists into one meta model to predict the stock market. The fund is backed by Union Square Ventures and the co-founder of Renaissance Technologies.

Define the exact aI Finance & Trading Tools workflow Numerai should support.

Compare it with closely related AI tools in the same category before committing.

Set review rules for accuracy, privacy, brand voice, compliance, and final approval.

Connect useful outputs to the wider stack instead of leaving them inside the AI tool.