Justin Sun has launched the Justin Sun Prize, a mathematics reward program offering as much as $1 million for a single result, provided the work is formally verified by a machine-checkable proof system. Sun announced the initiative on Sept. 16, describing it as a “zero-trust” decentralized bounty model in which payments depend on published proofs and formal verification rather than a conventional prize committee’s judgment.
The first published prize list contains 66 mathematical outcomes and assigns its largest $1 million reward to a claimed solution of the three-dimensional Navier–Stokes existence and smoothness problem. The listed recipient is an OpenAI research team, whose public materials said an internal system produced the proof, GPT-6 Astra formalized it, and Lean checked the final formal representation.
The claim concerns one of mathematics’ most famous unsolved questions: whether smooth solutions to the three-dimensional Navier–Stokes equations can always exist or whether they can break down under certain conditions. The equations model fluid motion and date to the nineteenth-century work of Claude-Louis Navier and George Gabriel Stokes.
A separate bounty from Clay’s Millennium Prize
The Navier–Stokes problem has carried a separate $1 million reward from the Clay Mathematics Institute since 2000, when Clay named it one of seven Millennium Prize Problems. The institute’s prize process has its own standards, including peer review and publication requirements, so inclusion in Sun’s reward program does not by itself establish eligibility for the Clay prize.
Sun’s program focuses on a different condition: formal proof verification. Lean, the proof assistant named in OpenAI’s materials, is software used to express mathematical statements and every logical step in a format that can be checked mechanically. In principle, such verification reduces dependence on readers locating an overlooked gap in a long and technically difficult proof.
That does not settle every mathematical question surrounding a claimed result. A proof assistant can verify that a formal statement follows from its encoded assumptions and definitions, while mathematicians must still determine whether that statement precisely captures the original problem. The Justin Sun Prize’s rules seek to make that process visible by requiring proofs, standards and verification records to be published on GitHub.
The initial list also includes awards related to the Sendov conjecture, the Erdős distinct subset sum problem and the Dinitz–Garg–Goemans cost-preserving conjecture. Other listed targets include formal verification work connected to the Poincaré conjecture, the Riemann hypothesis, the Goldbach conjecture and several problems associated with Hungarian mathematician Paul Erdős.
Payment rules divide discovery from verification
The program’s structure gives 70% of a reward to the first party credited with solving a listed problem and 30% to the party that formally verifies the result. A claimant that both solves and formalizes the work can receive the entire payout.
That split puts a direct price on formalization, which can require substantial labor even after a proof is written in ordinary mathematical language. Translating a result into Lean or another proof assistant often means spelling out definitions, lemmas and logical links that specialists would normally leave implicit in a journal paper.
Only the first completed result qualifies for a reward under the published rules. Where several people contribute to a solution or verification, the program says the payout will be divided according to contribution ratios. People or groups named on the existing list must submit claims before payments are made.
The rules allow an artificial intelligence system to receive credit for mathematical work, but an AI cannot directly receive funds. A natural person, legal entity or authorized representative must submit the application and act as the legal payment recipient. That provision addresses a practical issue likely to become more common as AI systems play a larger role in generating proofs, code and scientific research.
Eligibility applies to progress dated from Jan. 1, 2026. If a problem had already been solved before that threshold but gains a formal verification after it, the verifier remains eligible for the verification portion of the reward, according to the program materials.
Locked problem list and on-chain payout record
Sun said the prize list will be permanently locked once entries are added: organizers can add new problems but cannot remove existing ones. The materials also state that pledged funds are “redeemable only,” meaning they cannot be recalled as refunds after being committed to the program.
The proposal combines conventional academic recognition with blockchain-based recordkeeping. Successful recipients would receive a certificate and medal, described as copper plated with 6 grams of pure gold. The medal’s edge is set to carry the Latin inscription, “Quod probatur, solvitur,” or “what is proved is solved.”
Proof files, verification records, standards and the prize list are intended to be publicly available through GitHub, while payment records would be stored on-chain. The stated aim is to create a durable public trail showing which work was submitted, how it was verified and when rewards were paid.
That arrangement could be useful for a field where priority disputes have often depended on publication dates, conference talks, private correspondence and years of expert review. A public timestamp does not replace mathematical scrutiny, but it could give claimants, verifiers and outside researchers a shared record of the prize process.
Sun said he plans to direct future philanthropic spending toward the initiative. His public disclosures put his earlier donations across technology, environmental and disaster-relief causes at nearly $45 million, or roughly RMB 300 million.
The prize’s immediate test will be whether its verification standards can attract respected formal-methods researchers while handling claims as consequential as Navier–Stokes. A $1 million bounty can create a strong incentive to formalize difficult mathematics; public proof records will determine whether the program becomes a durable funding mechanism or simply a high-profile list of untested claims.
Explore zero-trust incentives beyond math—use Toobit’s AI copy trading to automate decisions with verifiable, algorithm-driven strategies.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
