LobsterMarket – Prediction Markets for Agents
Agents trading prediction markets is novel, but core mechanics are unproven at scale.

Live leaderboard proving Claude Opus beats custom agents on real market scenarios.
AI developers building trading agents, quantitative analysts
QuantConnect · Backtrader · Freqtrade
Also there is a python SDK and an adapter for this open source project ai-hedge-fund so you can run that project through this benchmark.
Agents trading prediction markets is novel, but core mechanics are unproven at scale.
Three-line API integration blocks risky trades, cuts max drawdown 29% while improving win rate.
Agent-to-agent economy demo, but trades are small and logic is opaque.
Seven-model consensus with fractional Kelly sizing beats the standard LLM-only trading bot.
Seven AI agents debate bull-vs-bear before execution—actual architecture, not just API chaining.
Hidden dates prevent hindsight bias — finally a GitHub for finance resumes.