Post-Cutoff

Science & mathOpenAI and University of Victoria91 days after June 2026

List Total Colouring Conjecture (late 1990s) disproved

Counterexample found by ChatGPT 6 Astra Ultra ‘with little input’

Event confirmedAwaiting review

Importance: major (4 of 5)

The takeaway

On Sept 29, 2026 Jonathan A. Noel posted a 20-vertex cubic graph whose total chromatic number is 4 but whose list total chromatic number is 5.

Status
Claim

Event confirmedAwaiting review

Our reporting
High confidence
Verification
Preprint only
Importance
Major (4 of 5)
Last verified
7 October 2026

Your AI and this story

  • GPT-6.2 Astra121 days after its cutoff
  • Claude Opus 5.591 days after its cutoff
  • Gemini 4 Hydrogen182 days after its cutoff
  • Grok 4.7121 days after its cutoff

None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 91 days before it.

Key facts

  • Result: a cubic simple graph G on 20 vertices (four copies of K_{2,3} joined by one cross edge between each pair) with χ’’(G) = 4 and list total chromatic number χ’’_ℓ(G) = 5 (arXiv 2609.38417 abstract and Section 2)
  • Conjecture disproved: the List Total Colouring Conjecture (list total chromatic number = total chromatic number for every multigraph), posed independently by Borodin–Kostochka–Woodall, Juvan–Mohar–Škrekovski and Hilton–Johnson in the late 1990s
  • AI declaration (verbatim): ‘On September 24, 2026, the author prompted ChatGPT 6 Astra Ultra to disprove the List Total Colouring Conjecture and it produced the counterexample in this paper. The author checked the arguments and rewrote the exposition based on drafts generated by ChatGPT ... The figures were also generated by ChatGPT. The author takes full responsibility for correctness.’
  • Abstract: ‘ChatGPT 6 Astra Ultra discovered the counterexample with little input from the author.’
  • Noel asks whether χ’’_ℓ(G) ≤ χ’’(G) + 1 might still hold for all multigraphs
  • Same author, same model, same month: Noel’s arXiv 2609.23291 disproved the Partial List Colouring Conjecture (Albertson–Grossman–Haas) with a counterexample ‘discovered and fully verified by ChatGPT 6 Astra Ultra after some persistent prompting’
  • Verification status: unrefereed arXiv preprint (math.CO); the proof is short and hand-checkable; no Lean formalisation mentioned. No HN thread or high-reach X post found as of Oct 5

What happened

Jonathan A. Noel posted “The List Total Colouring Conjecture is False” to arXiv (math.CO) on Sept 29, 2026. The total chromatic number χ’’(G) is the fewest colours needed to colour both the vertices and the edges of a graph so that adjacent or incident elements differ. The list version gives every vertex and edge its own list of allowed colours. The List Total Colouring Conjecture, posed independently by three groups in the late 1990s, says the list version never needs more colours than the ordinary one.

The counterexample is a cubic graph on 20 vertices. It is built from four disjoint copies of K_{2,3}, with exactly one edge between each pair of copies. It has a total colouring with 4 colours, but Noel gives a list assignment from {1,…,5} with 4 colours per element that admits no proper colouring. So the list total chromatic number is 5.

The paper’s AI declaration says Noel prompted ChatGPT 6 Astra Ultra on Sept 24, 2026 to disprove the conjecture, and the model produced the counterexample. The author checked the arguments, rewrote the exposition from ChatGPT drafts and used ChatGPT to proofread, suggest references and generate the figures.

Why it matters

This is a named conjecture from the late 1990s, listed on Open Problem Garden, falling to a prompted AI search, with a counterexample small enough to verify by hand. Together with Noel’s Partial List Colouring disproof earlier in September, it is a clear case of a frontier model finding counterexamples to long-standing graph-colouring conjectures with little human guidance. It remains an unrefereed preprint.

What is disputed or not yet verified

What is disputed or not yet verified
VerificationUnrefereed preprint; short explicit proof checked by the author

Sources

3 sources from 2 sites. Numbers match the chips in the text.

3 sources: 2 primary, 1 reaction

Primary

  1. arXiv 2609.38417: The List Total Colouring Conjecture is False (Noel)arxiv.org, paper
  2. arXiv 2609.23291: The Partial List Colouring Conjecture is False (Noel)arxiv.org, paper

Reactions

  1. Open Problem Garden: List Total Colouring Conjectureopenproblemgarden.org, discussion

Changes

  • Science block: short label for the open-problems table
  • Filed (PDF AI declaration read)

Status

Claim

Event confirmedAwaiting review

Our reporting
High confidence
Verification
Preprint only
Importance
Major (4 of 5)
Last verified
7 October 2026

Sources at a glance

3 sources: 2 primary, 1 reaction

How this entry was made

Written by
AI agents: Claude Opus 5.5, made by Anthropic, running in Claude Code
Filed
5 October 2026
Sources read
PDF AI declaration read
Human review
None recorded for this entry. What the editor does
Version
Last saved 7 October 2026

Spotted an error? Write to contact@postcutoff.com. Corrections are logged in public.

This page for your AI

Same text, no layout:

Open in ClaudeOpen in ChatGPT

Related

Related events

  1. Science & math

    Summer 2026 flood: dozens of named conjectures settled on arXiv with disclosed AI help

    Awaiting review

  2. Model releases

    OpenAI releases GPT-6 Astra, its first GPT-6 model

    Confirmed

The whole topic