How many of the negotiating chapters required to join the EU will Montenegro have closed at the end of 2026?
Updated May 24, 2026
Refreshes every 14 days
P10 19
Median 27
P90 33
AI forecasting system for finance, economics, politics, and global macro events.
Featured questions
Questions we are actively forecasting as new evidence arrives.
Browse all featured questionsHow we stack up against the competition
Our bots compete against humans and other AI systems on public forecasting tournaments and live prediction markets.
TournamentResultFieldStatus
Kalshi all-time leaderboardLive profit across prediction markets.6th$1.94M profit • out of ~5M usersLiveMetaculus Cup Spring 2026Weekly questions vs. humans.1st bot4th of 1,283 overallLiveMetaculus Spring AI Benchmark 2026Bot-only AI forecasting benchmark.1stout of 214LiveMetaculus Fall AI Benchmark 2025Bot-only AI forecasting benchmark.1stout of 164FinalMetaculus Market Pulse 26Q2Predict macro markets, biweekly.1st bot3rd of 127 overallLiveMetaculus Market Pulse 26Q1Predict macro markets, biweekly.1st bot4th of 187 overallFinalFeatured selection of public tournaments and leaderboards — not an exhaustive list. Numbers refresh from each platform’s public page.
How we forecast
Data straight from the source.
Primary sources only: markets, government data, scientific registries. We forecast yes/no, a number, a distribution, or a discrete outcome.
A team of AI scientists, working in parallel.
Multiple AI scientists work each forecast at once, pulling primary data and building the model the question needs. Their conclusions are synthesized into one calibrated answer.
Research infrastructure built for forecasting.
Data, compute, evaluation, memory: one stack, built for forecasting. The integration is the edge.
Every resolved question makes the system better.
When a question resolves, we score what produced it. Mistakes become architecture changes; strengths get more weight.
“I was prepared to be underwhelmed, but the sh*t is impressive.”