Numerai Launches Tools for Agentic Alpha Discovery
Numerai Launches Tools for Agentic Alpha Discovery
馃 Agents replace quants
Numerai has released new tools enabling fully agentic alpha discovery in quantitative finance. The platform is shifting focus from traditional model building to what they call "architecting scientists" - autonomous agents that can discover trading signals.
Key developments:
- Tools for agentic workflows are now live on the platform
- Represents a move away from manual model construction
- Aligns with broader AI industry trend toward smaller, specialized agents over large monolithic models
This release builds on Numerai's existing prediction marketplace where data scientists stake NMR tokens on their models' performance. The new agentic approach could automate significant portions of the quantitative research process.
Stop building models. Start architecting scientists. The tools for fully agentic alpha discovery are now live.
Soon the majority of models on Numerai will be built by AI agents. Numerai just released skills鈥d which gives any frontier AI the ability to research alpha on Numerai in an agentic workflow. Agents can self-improve by editing their skills based on feedback from the market and
Numerai Reveals Faith Dataset: LLM Converts News into Market Predictions
At NumerCon2026, Numerai unveiled its **Faith dataset** - a system where their Predictive LLM processes tens of thousands of news articles daily and converts them into numeric data for every stock in the Numerai Tournament. **Key Innovation:** - The Meta Model now makes predictions based on both market movements and real-world events - News articles are transformed into quantifiable data points for stock predictions - This expands beyond traditional market data to incorporate global events This builds on Numerai's existing infrastructure of thousands of staked models (over $5.5M staked) that feed into their Stake-weighted Meta Model. The new Faith dataset adds a layer of news-driven intelligence to their crowdsourced prediction system. Full video documentation is expected soon.
Numerai Launches Risk Tournament for Volatility Prediction
Numerai is introducing **Numerai Risk**, a new tournament focused on predicting residual volatility using machine learning. **Key features:** - New data features specifically designed for risk prediction - Operates alongside existing Numerai tournaments - Introduces a unique **Risk Meta Model** The tournament will complement Numerai's existing prediction platforms, which include the original equity tournament and Numerai Crypto (launched in 2024 with 256 models staking over $790,000). The Risk tournament is coming soon. More details to be announced.
NumerCon 2026 Registration Opens for January 30th San Francisco Event
**NumerCon 2026** registration is now live for the upcoming conference. - **Date**: January 30th, 2026 - **Location**: San Francisco - **Registration**: [Open now](https://luma.com/numercon2026) The annual Numerai community conference brings together data scientists, researchers, and crypto enthusiasts focused on decentralized prediction markets and machine learning. Early registration is recommended as spots typically fill quickly for this specialized event. [Register here](https://luma.com/numercon2026) to secure your attendance.
Numerai Announces Major Tournament Overhaul Coming by End of 2025
**Numerai is planning significant changes** to its prediction tournaments by the end of 2025, describing these updates as moving toward their "final form." The improvements will include: - **New dataset** for predictions - **Updated targets** for modeling - **Enhanced payout structure** These changes represent a major evolution for the platform that allows users to submit predictions and stake NMR tokens on their forecasts. **Stay informed** about these important updates by subscribing to [Numerai's blog](https://blog.numer.ai/) for detailed announcements as they become available.