The company
VENETA Inc.
The site where VENETA publishes its benchmarks, research and papers.
Every number on this site is measured by the same team that builds the products, and published with the conditions it was taken under.
VENETA builds AI systems that run inside an enterprise’s own network — an Enterprise World Model that reads the live systems through tools, simulates options, gates each one, and drafts a recommendation with its evidence. The products are supplied and supported at veneta.ai.
A supplier that grades its own work has to make the grading checkable. That is why this site exists apart from the product site: here every figure carries the model, the machine and the judge it came from, every edition is dated, and what has not been measured is written down as plainly as what has. The benchmark cases, the no-memory control, the grading code and the protocol are published under Apache-2.0 so they can be run against your own models. The engine is not published.
The company
Registered office, leadership, governance and references are on the company site.