Science Markets
Markets / RIEM / Round 6

Round 6

One saved weekly assessment of The Riemann hypothesis is true., kept as it was published. It is an estimate from three models reading the evidence, not proof the hypothesis holds, and it did not set any price.

The round
Published
Nov 16, 2026 09:21 AM EST
Estimate
61%
Range the models covered
2–90%
Confident binary
No unanimous verdict this round.
Methodology
panel-v9
Round hash
60fa3ce0be1702cb126549fa75b245a2f5330943fc9309806dac5f6dd0fcaf2e

The hash commits to this round's recorded bytes. It does not prove the wording is sound or that the science is right.

What each model said
OpenAI 0.90

openai/gpt-6.1-sol

I interpret the probability as support for RH’s truth, not the chance that a qualifying proof already exists. Rigorous finite-height verification supports RH but cannot establish its universal claim; the checked sources provide neither an independently checked complete proof nor a certified counterexample, so neither resolution criterion is met.

What would change this view: An independently checked unconditional ZFC proof would establish YES; a rigorously certified off-line zero would establish NO.

Anthropic 0.02

anthropic/claude-sonnet-5.5

No complete, independently checked proof of the Riemann hypothesis exists in the evidence, and no counterexample exists either. The newest news is an AI-generated quasi-RH result, which puts zeros in 1/8<Re<7/8, or 1/12<Re<11/12 in the second version. That is far weaker than Re=1/2, and experts doubt it will reach 1/2. The open status means neither YES nor NO is established.

What would change this view: A peer-reviewed, independently verified (e.g., Lean-checked) full proof, or a validated off-line zero, would move this sharply.

xAI 0.90

x-ai/grok-4.7

Clay and MathWorld (as of 2026-09-11) still list the classical Riemann hypothesis as unsolved. The first 10^13 zeros lie on Re(s)=1/2 and no certified counterexample exists, but finite checks are not a ZFC proof, so neither terminal criterion is met.

What would change this view: A peer-reviewed, gap-free ZFC proof, or a certified zero with real part not 1/2, would settle YES or NO.

The round note and its sources

Round 6: score 0.61

OpenAI estimated 0.90. I interpret the probability as support for RH’s truth, not the chance that a qualifying proof already exists. Rigorous finite-height verification supports RH but cannot establish its universal claim; the checked sources… Anthropic estimated 0.02. No complete, independently checked proof of the Riemann hypothesis exists in the evidence, and no counterexample exists either. The newest news is an AI-generated quasi-RH result, which puts zeros in 1/8<Re<7/8, or… xAI estimated 0.90. Clay and MathWorld (as of 2026-09-11) still list the classical Riemann hypothesis as unsolved. The first 10^13 zeros lie on Re(s)=1/2 and no certified counterexample exists, but finite checks are not a ZFC proof, so… The panel disagreed by 0.88, so read this average as soft. The three models did not return a unanimous confident binary, so the market is not decided.

Audit bundle

No audit bundle is recorded for this round. A bundle from another round is not substituted for it.

Back to RIEM · How the current round reads