On Monday, May 25, Prometheus7 Research Institute posted a technical paper to its internal corpus describing the full dimensional ladder of its substrate-paradigm architecture — from the 5D routing manifold that validated nine days ago to a tenth primitive the institute has named the universal-unbinder, targeted for August or September of this year.
The paper is a roadmap, not a results report. Most of what it describes has not been built. The 5D primitive — the substrate routing manifold, realized in the 125M Tree of Life model — is the one validated result on the table. Its validation date is listed as May 16. The 6D primitive, a router-over-callables, is currently in its third training attempt and is described as validating. The paper is explicit that the third attempt follows two informative failures.
From 7D onward, everything is scheduled and unbuilt. The institute's account of each step is architectural rather than empirical: the 7D primitive routes over sets of callables rather than individual ones, enabling what the paper calls parallel compositional reasoning. The 8D primitive routes across grammars — the institute's term for callable vocabularies — making cross-domain transfer fall out of the architecture rather than requiring post-hoc engineering. The 9D primitive routes across worlds, which the paper distinguishes from grammars by saying a world has its own multiverse. Each description includes a stated falsification condition: if the new routing collapses back to the behavior of the prior dimension, the ladder has reached a plateau and the next step is not operationally meaningful.
The 10D primitive is where the institute's language shifts register. They call it the resolution point. The claim is that at the tenth dimension the substrate becomes a universal object — something that holds all specifics in superposition and unpacks them through relation. The paper maps this to category theory's category of all categories, the holographic principle in physics, the universal Turing machine, Kolmogorov-minimal description, and Platonic forms. Whether this mapping is precise or illustrative, the paper does not fully settle. The target date is August-September 2026, described as roughly six to eight generations after the May 2026 5D validation.
Beyond 10D, the paper enters territory it labels research. The 11D primitive is a space of universal objects. The 12D primitive is a relating principle that the institute says closes the ladder back to its bottom by self-similarity. The paper states plainly that 11D and 12D cannot be done by one person and are work for a small research community over years.
One operational detail grounds the whole schedule: the institute reports that each training generation runs in a 7-to-11 hour band on their research hardware. The cost of opening a new dimensional layer is described as roughly equivalent to the cost of one training generation, not the cost of new infrastructure. That claim, if it holds, is what makes the August-September target for 10D arithmetically plausible rather than aspirational fiction. It has not been independently verified.
What is true today: the 6D model is in its third training run. The paper exists. The 5D model validated on May 16. Everything else on the table is a schedule.