Will any frontier lab publish its flagship's absolute layer count? — Contested · resolves 2026-11-01

The debate: Geoffrey Irving, chief scientist at the UK AI Security Institute, argues that there is no need to wait for legislation and that an independent person could review frontier model architecture now, and his own test requires an absolute figure, the computational-graph depth or layer count of a current flagship model.

Why it matters: OpenAI's Pachocki offered only a relative figure on September 2, within twice GPT-4, so the yardstick and the reading do not match, and if none of the three frontier labs publishes an absolute figure the prescription cannot be carried out under the current disclosure regime.

Who is on each side: Irving bets that independent review is possible today; our 2026-09-06 column rated that claim contested.

This card is a yes-or-no question, not a bet on either side: by 2026-11-01, will OpenAI, Anthropic or Google DeepMind publish the absolute computational-graph depth or layer count of a current flagship model (or an equivalent verifiable scale), in official documentation, a system card or a report by a named independent party? If one does, Irving's prescription can be carried out and our column's contested rating moves to supported; if none does, the conclusion that his criterion cannot be met under the current disclosure regime is confirmed.

Sources still being linked up — the node(s) behind this call don't have a clickable original link resolved yet. Logged for follow-up.

Only the labs' own system cards, model cards and technical documents, and signed reports from named independent evaluators such as the UK AI Security Institute, count; media reports do not. Publication of an absolute depth or layer count (or an equivalent verifiable scale) by any of the three before the due date answers yes; none answers no. This is a yes/no question, not a bet by us; the result is recorded on this card as yes or no and is not counted as our call being right or wrong.

What would overturn it

If it emerges that the existing agreements of the UK AI Security Institute or other invited third parties already cover architecture details, so that independent parties can already see the architecture, Irving's premise that the labs have not done this is false and the column's rating changes to refuted. If the original text or source material behind The Information's report becomes public and shows that the technique it named does suppress the model's chain of thought, the factual basis of the whole dispute has to be reread and this card is void.