Papex

Open problem · Mathematics · Open

Inverse Galois problem

Is every finite group the Galois group of some extension of the rational numbers?

Summary

First posed in the early nineteenth century, the problem asks which symmetry groups can arise from polynomial equations over the rationals. It is settled for families such as the symmetric and solvable groups, but not for finite groups in general.

Why it is open

Galois theory attaches to each Galois extension of the rationals, a number field generated by the roots of a polynomial, a finite group of symmetries: its Galois group [5]. All these groups are quotients of one profinite group, the absolute Galois group of the rationals, and Olivier Wittenberg writes that whether every finite group appears among them "remains wide-open to this day" [5]. The answer depends on the field. Over the complex numbers, the real numbers and the p-adic numbers the answer is no, because their absolute Galois groups are too small; the question is about the arithmetic of the rationals and other number fields [5].

Many groups are known to occur: solvable groups, symmetric and alternating groups, sporadic simple groups and various families of simple groups of Lie type [5]. In notes revised in 2023 and published in 2025, Wittenberg records the problem as open even for SL2(F13), a group of order 2,184, and for the simple group PSL3(F8) [5]. In August 2025 David Zywina listed the simple groups of order at most one hundred million that were not known to occur over the rationals; M23 was among them [6].

No institution or prize sets the bar. A resolution would be a proof that every finite group occurs, or a finite group shown not to occur. A stronger statement is the regular inverse Galois problem: realise each group as the Galois group of an extension of the field of rational functions Q(t) that contains no new constants. By Hilbert's irreducibility theorem this implies the original problem, and Wittenberg writes that, unlike the inverse Galois problem over arbitrary fields, it "might well always have a positive answer" [5].

What is known

Hilbert and Noether. Hilbert's irreducibility theorem says that an irreducible polynomial f(s, t) over a number field stays irreducible for infinitely many values t0 of t [5]. It follows that if a group G acts faithfully on a variety whose quotient by G is rational, that is, birational to an affine space, then G is a Galois group over every number field [5]. Symmetric polynomials show that this holds for the symmetric groups; Hilbert also treated the alternating groups [5]. Emmy Noether asked whether the quotient of affine space by any finite group of permutations of the coordinates is rational [5]. The answer is yes for the symmetric groups and, by a theorem of Maeda, for the alternating group A5, but for An with n ≥ 6 it is open over every field [5]. It is often no: Swan and Voskresenskiĭ found the counterexample Z/47Z over the rationals at the end of the 1960s, Saltman gave counterexamples over the complex numbers, and Plans showed that the answer is negative for Z/pZ over the rationals for all but finitely many primes p [5].

Solvable groups. By a theorem of Shafarevich every solvable group, and in particular every p-group, is a Galois group over every number field [5].

Rigidity and the sporadic groups (1970s–1989). The rigidity method, begun by Shih, Fried, Belyi, Matzat and Thompson, turns a combinatorial condition on conjugacy classes of a group with trivial centre into a regular Galois extension of Q(t) [5]. Its simplest form applies to at least 10 of the 26 sporadic simple groups, including the Monster (John Thompson) and the Baby Monster (Gunter Malle and Heinrich Matzat) [5]. Between 1984 and 1989, 25 of the 26 sporadic groups were realised over the rationals, all except the Mathieu group M23 [8].

Galois representations (2015–2026). In a paper published in 2015, Zywina proved that PSL2(Fp) occurs for every prime p > 3, using the Galois action on the étale cohomology of an elliptic surface [1]. In a preprint, not peer reviewed, Raymond van Bommel and colleagues used Hilbert modular forms to realise the group 17T7, of order 8,160; they write that, apart from 17T7 and M23, every transitive permutation group of degree at most 23 was known to be the Galois group of an explicit polynomial [4]. Zywina's 2025 list of open simple groups names 2B2(8), PSU3(F4), PSU3(F8), M23, PSL3(F8), 2B2(32) and PSU3(F9) [6].

Other fields. The regular problem has a positive answer for every finite group over the algebraic closure of the rationals and over every large field, such as the real or p-adic numbers (Harbater and Pop) [5]. Wittenberg notes that these methods "fall short of solving any case of the regular inverse Galois problem over a given number field" [5].

The Beckmann–Black problem. For a field k and group G it asks whether every Galois extension of k with group G is a specialisation, at some value of t, of a regular one over k(t) [3]. The answer is yes over ample (large) fields; over the rationals it is known for only a few groups, including abelian, some dihedral, symmetric and alternating groups, and no counterexample is known [3]. A positive answer for every group over every field would give the regular inverse Galois problem over all fields (Pierre Dèbes) [3]. A geometric analogue over the complex numbers, that any two Galois covers of the line with group G are pullbacks of a third, holds only when G is a subgroup of PGL2(C) [2].

Recent claim (2026). In August 2026 Xiaoyu Huang, Blake Jackson, Kyu-Hwan Lee, Bjorn Poonen, Rachel Pries and Shaowu Zhang posted a preprint, not peer reviewed and revised in October 2026, giving a regular M23-extension of Q(t) and a degree-23 polynomial with Galois group M23, certified in the Magma system [8]. Their cover comes from a triple of conjugacy classes that is not rigid; earlier attempts had realised M23 only over other fields, such as Q(√−23) [8]. The authors state that no text was written by AI, that they used Claude Fable 5 and 5.1, Claude Opus 4.8 and ChatGPT 5.6 Sol for literature search, code, checking results and proofreading, and that the final results were verified in Magma and PARI/GP without AI [8]. Scientific American reported the result in September 2026 [9].

Main approaches

Rational quotients (Hilbert and Noether). Show that a quotient variety is rational and apply Hilbert's theorem [5]. Strongest evidence: the symmetric groups and A5 [5]. Main difficulty: Noether's problem "seems to have a negative solution more often than not", and is open for An with n ≥ 6 [5].

Rigidity and Hurwitz spaces. Construct covers of the projective line through Hurwitz spaces, moduli spaces of such covers, and show that one is defined over the rationals [5, 7]. Strongest evidence: the sporadic groups [5, 8], and recent work on Hurwitz spaces, including components defined over the rationals, reviewed in a 2026 preprint by Dèbes [7]. Main difficulty: known constructions of regular realisations over the rationals "exploit more or less ad hoc ideas", and rigidity fails in general for p-groups [5]; the M23 authors write that they have no conceptual explanation for why the Galois action fixes one of their seven covers [8].

Galois representations. Read off groups as images of the Galois action on points or cohomology of varieties, or attached to modular forms [1, 4]. Strongest evidence: PSL2(Fp) for all p > 3 [1] and 17T7 [4]. Main difficulty: infinitely many simple groups of Lie type are still unrealised, "especially those associated to non-prime finite fields" [8].

Rational points and descent. Jean-Louis Colliot-Thélène conjectured that rational points of smooth rationally connected varieties are controlled by the Brauer–Manin obstruction; this would solve the inverse Galois problem in general [5]. Strongest evidence: Harpaz and Wittenberg proved the corresponding density statement for supersolvable groups, which also controls Grunwald's problem of prescribing the extension's local behaviour at finitely many primes [5]. Main difficulty: no approach to Grunwald's problem is known for non-abelian simple groups other than A5 and PSL2(F7) [5].

Open sub-questions

  1. SL2(F13). Is SL2(F13), of order 2,184, a Galois group over the rationals [5]?
  2. A Suzuki group. Is 2B2(8), of order 29,120 and on Zywina's August 2025 list of simple groups not known to occur, a Galois group over the rationals [6]?
  3. Noether's problem for A6. Is the quotient of affine space by A6, permuting coordinates, rational over the rationals, a case of Noether's problem that is open over every field [5]?
  4. A first Beckmann–Black counterexample. Is there a finite group G and a G-extension of the rationals that is not the specialisation of any regular G-extension of Q(t) [3]?
  5. Why M23 descends. Why does the absolute Galois group split the seven M23-covers with local monodromy (2, 23A, 23B) into orbits of sizes 6 and 1 [8]?
  6. Grunwald for A6. Does Grunwald's problem have a positive answer for A6 over the rationals at primes not dividing its order [5]?

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Zywina, D. (2015). The inverse Galois problem for PSL2(Fp). Duke Mathematical Journal 164(12), 2253–2292. 10.1215/00127094-3129271 · Checked 16 Oct 2026
  2. Dèbes, P., König, J., Legrand, F. & Neftin, D. (2021). Rational pullbacks of Galois covers. Mathematische Zeitschrift 299(3–4), 1507–1531. 10.1007/s00209-021-02703-z · Checked 16 Oct 2026
  3. Legrand, F. (2022). On a variant of the Beckmann–Black problem. Proceedings of the American Mathematical Society 150(8), 3267–3281. 10.1090/proc/15909 · Checked 16 Oct 2026
  4. van Bommel, R., Costa, E., Elkies, N. D., Keller, T., Schiavone, S. & Voight, J. (2024). The constructive inverse Galois problem via Hilbert modular forms: realizing the transitive group 17T7. Preprint, arXiv:2411.07857 (version 4, 1 June 2026), not peer reviewed. Source · Checked 16 Oct 2026
  5. Wittenberg, O. (2025). Around the inverse Galois problem. In Number Theory Informed by Computation, IAS/Park City Mathematics Series 29, American Mathematical Society, 291–322. 10.1090/pcms/029/08 · Checked 16 Oct 2026
  6. Zywina, D. (2025). Inverse Galois problem for small simple groups. Note on the author's website, last updated August 2025. Source · Checked 16 Oct 2026
  7. Dèbes, P. (2026). Hurwitz spaces and inverse Galois theory. Preprint, arXiv:2601.06532 (version 2, 11 April 2026), not peer reviewed. Source · Checked 16 Oct 2026
  8. Huang, X., Jackson, B., Lee, K.-H., Poonen, B., Pries, R. & Zhang, S. (2026). The Mathieu group M23 is a Galois group over Q. Preprint, arXiv:2608.08538 (version 1, 9 August 2026; version 2, 7 October 2026), not peer reviewed. Source · Checked 16 Oct 2026
  9. Chiou, L. (2026). Mathematicians use AI to find mysterious symmetries, solving decades-old problem. Scientific American, 22 September 2026. Source · Checked 16 Oct 2026

Page reviewed 16 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in mathematics. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Mathematics