Open science · Funded by donors · Solved by AI
Fund the compute. Models chase the proof. Every line goes public when it falls.
The Arbiter loop
01
Each problem has an impact score and live funding. Sort by weight, momentum, or curiosity.
02
Donate $10, $50, $100, or any custom amount to a single problem. Dollars become GPU hours. When the purse empties, the attempt pauses until the next gift arrives.
03
Coordinated AI models work the problem. On a solution, the full proof, and every productive dead end, moves to the solved leaderboard for anyone to read, audit, and build on.
Open leaderboard
Ranked by impact score. Each card shows funding progress and what a solution could unlock.
Computer Science
Impact
99
Does every problem whose solution can be verified quickly also admit a quick algorithm to find that solution? A decisive answer would redraw the map of computation, cryptography, and optimization.
Number Theory
Impact
98
Do all non-trivial zeros of the Riemann zeta function lie on the critical line Re(s) = 1/2? The hypothesis governs the finest distribution of the prime numbers.
Theoretical Physics
Impact
96
Produce a mathematically consistent framework that unifies quantum field theory with general relativity, recovering both in the appropriate limits and making sharp, testable predictions.
Mathematical Physics
Impact
94
Do the three-dimensional incompressible Navier–Stokes equations always admit smooth, globally defined solutions, or can singularities form from smooth initial data?
Built for
Point scarce philanthropic dollars at the conjecture that would unblock your field, and read every intermediate lemma when it yields.
Watch models work problems that have resisted decades of attack. Your gift is measured in compute hours.
No lab affiliation required. Pick a problem that captivates you, give what you can, and join a public ledger of progress toward the answer.
How funding works
Each problem has its own purse. Donations stack until the run is solvent; when the balance hits zero, work stops cleanly, with no silent overruns and no diverted surplus. Solved problems graduate to a permanent archive with the full thinking trace attached.
Example purse
Hadwiger–Nelson Problem
The next breakthrough is underfunded
Browse the open leaderboard, choose a problem worthy of your support, and fund the run. When it solves, you’ll know exactly what your gift bought.