This interactive needs WebGL2 with floating-point render targets, which this browser does not provide.
AI Mathematics · Number theory
Riemann’s Line, 1859 → 2026 Landscape · The zeta landscape
Every zero of the zeta function found so far is a well on one straight line. This is the ground they sit in.
What you are looking at The zeta function takes a complex number s = σ + it and returns another. The ground is the complex plane: across, the real part σ from −11.5 to 12.5, twelve units either side of the line you fly along; along, the imaginary part t, the height up the strip, which is the direction you fly. Over each point the surface rises or falls with the size of ζ(s), and its colour gives the direction of ζ(s) in the complex plane: each hue is one direction, and the hues turn once around every zero.
height = 1.4 · tanh( ln(1 + |ζ(s)|) / 1.4 )
colour = arg ζ(s)
The height grows with |ζ| and with nothing else. It is exactly 0 where ζ is zero, and near a zero it rises in proportion to |ζ|, so every zero is a true cone. Far from the zeros it grows like the logarithm of |ζ|, and the tanh levels it off toward 1.4: over the pole, and over the left half-plane, where ζ grows enormous (at height 1,000, σ = −1, |ζ| is about 1,900). The terrace on the left is that ceiling, not a place where ζ stops growing. The flat plain on the right is where |ζ| is close to 1. The relief slider stretches the drawing only. The left half-plane is dimmed outward from the strip so the strip reads first; that shading is presentation and encodes nothing.
The thin dark lines mark each doubling of |ζ|; the optional light ones, every 30 degrees of phase. Light comes from ahead; the darkening inside the wells follows the same height and adds nothing new.
Wells, a spire, a line A zero is a point where ζ(s) = 0: the surface falls into a well and all the colours meet at its bottom. The bright hairline is real part ½, the critical line, between the two fainter edges of the critical strip, 0 < σ < 1. Near the start you pass the spire of the pole at s = 1, where ζ is infinite, and, on the left at t = 0, the trivial zeros at −2, −4, −6, −8 and −10, which are understood. Far to the right ζ settles to 1, the flat plain; far to the left it is astronomically large (at height 1,000 and σ = −11.5, about 10²⁶), and the terrace is the ceiling of the height. The gold glints mark the zeros from Andrew Odlyzko’s table; the wells beneath them are computed independently, here, from ζ itself.
The Riemann Hypothesis says every other zero lies on the critical line. It is the statement that each of these wells, all the way up, sits on the bright hairline and nowhere else in the strip. Near height 7005 two wells almost touch: Lehmer’s pair, zeros 6,709 and 6,710, only 0.038 apart, the kind of near miss that would let a zero slip off the line if the hypothesis were false.
Curiosities Places worth a closer look. Choose one to fly there; turn on Curiosities in the controls, or press c , to see their letters on the ground. The same place keeps its letter in all three views.
The announcement On 20 November 2026, Levent Alpöge announced that, working with Claude, he had found a proof of the Riemann Hypothesis, extending to the Generalized Riemann Hypothesis for Dirichlet L-functions (his note ). He released it only as a formalization in the Lean proof assistant, 175,187 files and about 17.4 million lines, and wrote that a readable account would follow. The same day, Sahibzada Allahyar posted an expository reconstruction of the argument.
Status This page does not check the proof. As of 21 November 2026 the claim is unverified: an independent build on a public server whose operator is not named reported on 20 November that the whole project compiles with no failures and that its final theorems rest only on Lean’s three standard axioms (a fresh kernel replay was skipped; the audit ), but no named mathematician or peer review has confirmed it; the reconstruction states that the full dependency tree and the axioms of the final declaration have not been independently checked for it. What the page shows is classical and independent of the claim. The zeros marked are numerical values from Odlyzko’s tables (the first 10,000, to height 9,878), and the hypothesis has been verified numerically for every zero up to height 3 × 10¹² (Platt and Trudgian, 2021). A landscape this size shows no more than the first 10,000.
How it is computed The surface is ζ computed live on your graphics card by Euler–Maclaurin summation, the formula of the companion page with N = |s|/π + 3 terms, in 32-bit arithmetic: each block of heights is referred to a nearby height whose phases are prepared in 64-bit, so the large angles t·ln n never lose precision. Within the strip and near it, checked against 64-bit values at 30 points (σ from −1 to 2, heights 1 to 9,878), the relative error stays below 6 × 10⁻⁶ wherever |ζ| exceeds 10⁻³, the phase within 0.0003 degrees, and at the zeros themselves the value is within 10⁻⁷ of 0. Out on the wide plane: on the right, below 10⁻⁷. On the left near t = 0, where the direct sum would cancel ten orders of magnitude, ζ is computed through the functional equation instead and stays within 2 × 10⁻⁶. Higher up on the far left, under the ceiling, 32-bit arithmetic gives the colour to within 0.15 degrees (relative error up to 3 × 10⁻³). Columns are packed most densely in the critical strip (1/256 apart) and spread to about 1/30 at the far edges, where the ground is flat. Flight stops at height 9,860, the end of the zero table.