The abc Conjecture

Pick two whole numbers with no common factor — say a = 5 and b = 27 — and add them: c = 32. Now look at the primes hiding inside all three. The number 5 is prime; 27 is $3^{3}$; 32 is $2^{5}$. Strip every repeat and multiply the distinct primes together: 5 × 3 × 2 = 30. That product is called the radical of abc, written rad(abc).

Here is the strange part. Even though 27 and 32 are built from towering powers of small primes, their radical, 30, is barely smaller than c = 32. The abc conjecture says this is no accident: for coprime a + b = c, the sum c is almost never much larger than rad(abc). Triples where c dwarfs its radical are extraordinarily rare.

It sounds like a curiosity about addition. It is in fact one of the most powerful unproven statements in mathematics — a single rule that, if true, would settle a long list of famous problems at once.

For any coprime a + b = c we measure how "extreme" the triple is with its quality: q = log c / log rad(abc). When c barely beats its radical, q sits just below 1. When c is much larger than rad(abc), q climbs above 1 — and the abc conjecture says triples with q above any fixed bound bigger than 1 are finite in number.

<p class="hint">{{hint}}</p>
<div class="inputs">
  <label>a <input id="a" type="number" min="1" value="5"></label>
  <label>b <input id="b" type="number" min="1" value="27"></label>
  <button id="go" type="button">{{btn_compute}}</button>
</div>
<div id="out" class="out"></div>
<div class="btns">
  <button id="search" type="button">{{btn_search}}</button>
  <button id="reset" type="button" class="ghost">{{btn_reset}}</button>
</div>
<div id="board" class="board"></div>

* { box-sizing: border-box; }
body { font-family: system-ui, sans-serif; color: #222; margin: 0; }
.hint { font-size: .9rem; color: #444; margin: 0 0 .7rem; line-height: 1.45; }
.inputs { display: flex; gap: .6rem; align-items: center; flex-wrap: wrap; margin: .4rem 0; }
label { font: 600 14px system-ui, sans-serif; color: #1d3557; }
input { width: 84px; font: 600 15px ui-monospace, monospace; padding: .35rem .5rem;
        border: 1px solid #cdd9e3; border-radius: 8px; margin-left: .25rem; }
.out { background: #e8eef3; border: 1px solid #cdd9e3; border-radius: 8px; padding: .7rem .9rem;
       margin: .6rem 0; font: 600 14px ui-monospace, monospace; color: #1d3557; min-height: 1.4em;
       line-height: 1.6; }
.out.bad { background: #fdecee; border-color: #f3b6bd; color: #c92f3c; }
.out b { color: #0a7d33; }
.btns { display: flex; gap: .5rem; flex-wrap: wrap; margin: .3rem 0 .2rem; }
button { font: 600 14px system-ui, sans-serif; padding: .45rem .9rem; border: 1px solid #1d3557;
         background: #1d3557; color: #fff; border-radius: 8px; cursor: pointer; }
button.ghost { background: #fff; color: #1d3557; }
.board { margin-top: .5rem; }
.row { display: grid; grid-template-columns: 1.6fr 1fr 1fr 1fr; gap: 6px; font: 600 13px ui-monospace, monospace;
       padding: .35rem .5rem; border-radius: 6px; }
.row.head { background: #1d3557; color: #fff; }
.row.item { background: #f1f4f7; color: #1d3557; }
.row.item:nth-child(even) { background: #e8eef3; }
.q { font-weight: 700; color: #c92f3c; }

// Code not found

Type your own a and b to see c, the radical, and the quality. Then press Search to sweep small coprime pairs and surface the highest-quality triples found — the famous record holder 2 + $3^{10}$·109 = $23^{5}$ with q ≈ 1.63 is exceptionally hard to beat. Notice the asymmetry: computing q for one triple is instant, but finding record triples means combing through a vast space, and proving they thin out forever is another matter entirely.

How hard is the abc conjecture? Not the checking — the proving.

• Verifying a single triple is trivial: factor a, b and c, take the distinct primes, divide the logarithms. A computer does it instantly, and projects like ABC@Home have tested billions of triples — every one obeys the pattern.
• The conjecture itself is OPEN. Stated independently by Joseph Oesterlé and David Masser in 1985, it claims that for every $\varepsilon > 0$ only finitely many coprime triples have $c > \mathrm{rad}(abc)^{1+\varepsilon}$. No amount of computation can confirm "finitely many" — that needs a proof.
• The disputed proof. In 2012 the mathematician Shinichi Mochizuki released a ~500-page proof built on a new framework he calls Inter-universal Teichmüller (IUT) theory. It was published in 2021, but a key step (Corollary 3.12) was challenged in 2018 by Peter Scholze and Jakob Stix, who argued it contains a gap. The dispute is unresolved, so the wider community does not regard abc as proved.
• So abc sits in limbo: overwhelmingly supported by data, claimed-but-contested in theory, and still listed among the great open problems of number theory.

This is a different flavor of hardness from P vs NP. There the obstacle is computational; here it is a question of truth and consensus — a proof can exist on paper yet fail to convince.

Mathematicians care about abc because it is a master key: prove it once and a remarkable list of hard results follows almost for free.

• Fermat's Last Theorem (for large exponents): abc gives a short, near-immediate proof that $a^{n}$ + $b^{n}$ = $c^{n}$ has no solutions once n is large enough.
• The Mordell conjecture / Roth's theorem: deep statements about rational points on curves and how well algebraic numbers can be approximated fall out as consequences.
• The Szpiro conjecture and elliptic curves: abc is essentially equivalent to a bound on the conductor and discriminant of elliptic curves.
• Bounds on Diophantine equations: a uniform handle on how large solutions of Diophantine equations can get.

That is why a clean, accepted proof of abc would be a landmark: it concentrates the difficulty of dozens of problems into a single statement about three coprime numbers.

The abc conjecture distills an enormous amount of mathematics into a single sentence: when coprime numbers add, the sum rarely escapes the primes hiding inside them. Billions of examples agree; a proof exists on paper but not yet in the community's consensus.

So the conjecture stands as a vivid reminder that checking and proving are worlds apart. You can verify any triple in a heartbeat, watch the pattern hold every single time — and still not know, with certainty, that it holds forever. Like P vs NP, abc marks the edge of what we can compute and the deeper edge of what we can prove.