Claimed AI-led

Shelah's eventual categoricity conjecture for abstract elementary classes

An OpenAI manuscript claims a ZFC proof of Shelah's eventual categoricity conjecture, that an abstract elementary class categorical in one sufficiently large cardinal is categorical in every larger one, a statement for which Christian Espíndola has had a claimed ZFC proof posted since 2019.

Lab: OpenAI
Model: Unreleased internal OpenAI model, not named in the release
Field: Mathematics
Date: 2026-10-06

Sources

What was found

The main theorem gives, for every infinite $\lambda$ , a threshold $\mu(\lambda)$ such that any abstract elementary class with Löwenheim-Skolem number at most $\lambda$ that is categorical in one cardinal at or above $\mu(\lambda)$ is categorical in all of them, with no amalgamation, joint embedding, tameness or no-maximal-models hypothesis and no restriction on the starting cardinal. It is the nonelementary analogue of Morley's theorem and covers $L_{\omega_1,\omega}$ sentences in countable languages. A companion manuscript shows the threshold cannot be prescribed in general: under the continuum hypothesis there is a class with Löwenheim-Skolem number $\aleph_0$ categorical on a tail of cardinals but not at $\beth_{\omega_2}$ , so the prescribed-threshold form is not provable in ZFC if ZFC is consistent. Earlier positive results needed large cardinals (Makkai and Shelah; Boney), tameness with amalgamation (Grossberg and VanDieren), or special classes (Vasey for universal classes; Shelah and Vasey for excellent ones).

Novelty check

Read the main manuscript's introduction and bibliography at commit fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb on 2026-10-12. The live novelty question is Christian Espíndola's 'A proof of Shelah's eventual categoricity conjecture and an extension to accessible categories with directed colimits', arXiv:1906.09169, first posted in 2019 and in a seventh version since April 2022, which states a ZFC proof for abstract elementary classes. The OpenAI manuscript cites it, calls it a separate public claim, and says none of its interfaces are imported from it; it does not say whether that proof is correct. Whether this is the first proof or a second one cannot be settled until someone qualified checks one or both.

Caveats and known objections

The family's Lean formalization covers only the companion CH obstruction, not the eventual categoricity theorem: its scope document says the separate eventual-categoricity theorem is not included. The main manuscript runs to fourteen sections and roughly 460,000 characters of source, and builds its own amalgamation, independence calculus and geometry from the axioms rather than importing them, which is exactly where earlier attempts in this area have had gaps; the manuscript itself notes Boney and Vasey's identification of a gap in Shelah's own Chapter IV. Nobody outside OpenAI is on record as having read it. Autonomy is graded ai-led for the same reasons as the other families split from this release.

The open problem

MathDBShelah's eventual categoricity conjecture

389081

Nothing mechanically. It records what a problem says, what is known about it, and the standing of any claimed solution.

Also recorded at

vibemathedshelah-eventual-categoricity-conjecture

Nothing mechanically. It is a parallel listing of the same result, carrying its own verification label rather than an independent check.

Nobody outside the lab has checked this yet.

Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.

Submit a check Challenge the grade

Or on GitHub: submit a check challenge the grade send a correction or send a pull request

History of this entry

2026-10-16CorrectedRecorded the vibemathed record for this result and the MathDB record for the conjecture as registrations. No grade moved.
2026-10-12AddedEntered the registry graded Claimed and AI-led.

Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.

Community discussion

Hacker NewsSharing AI progress in mathematics↗

Graded claimed for verification and ai-led for autonomy. What these mean.

Cite this entry

Plain text

whataifound.org. (2026). Shelah's eventual categoricity conjecture for abstract elementary classes. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-06-shelah-eventual-categoricity

BibTeX

@misc{whataifound-openai-2026-categoricity,
  title        = {Shelah's eventual categoricity conjecture for abstract elementary classes},
  author       = {{whataifound.org}},
  year         = {2026},
  howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
  note         = {Result by OpenAI. Verification: Claimed. Autonomy: AI-led.},
  url          = {https://whataifound.org/finding/2026-10-06-shelah-eventual-categoricity}
}

Related findings

Author verifiedEquivalence of generic stability notions for Keisler measuresIndependent · 2026-08-25 Author verifiedCounterexample to the stable forking conjectureIndependent · 2026-08-31 Author verifiedSOP₂ and SOP₃ theories shown to coincideIndependent · 2026-08-13

model-theory logic set-theory argument

← All mathematics findings in the registry