Another scenario shouldn't cost another week.

We publish our benchmarks in full, including what they do not show. There are two so far, and more in progress.

Google OR-Tools. A full ten-year plan, benchmarked against Google OR-Tools on the same orebody, the same constraints and the same assumptions. Cognomine returned a plan in 25 seconds. Google OR-Tools, given the same variables and the same constraint set, was stopped unconverged at 12 hours. The Cognomine plan carried $273M more discounted NPV, about 16% higher, on $149.5M less equipment capital, around 59% less, with a peak fleet of 2 diggers and 17 trucks against 5 diggers and 41 trucks.

MineLib. A published third-party solution on a public dataset, replayed under the physical constraints a real operation works under. It measures how much of a plan's stated value survives contact with fleet, haulage and drill-and-blast timing. Most of it does not.

Further benchmarks on additional public datasets are in progress and will be published as they complete.

Pit selection, block scheduling and haulage usually mean three steps in three tools. Keeping them aligned as assumptions move eats a good share of any study.

Cognomine solves all three in a single pass. Ultimate pit selection decides what is economic to extract, the block schedule decides what gets mined when, and haulage decides how the material moves and what fleet that needs. Because they resolve together rather than one after another, there is one set of inputs to prepare and one result to defend.

Strategic mine planning has always been limited by what a solver can accept simultaneously. To get a long-horizon answer back in reasonable time, operating detail gets dropped, and even then the run still takes hours or days. So a study that needs many scenarios goes out with a couple.

Cognomine encodes more of that detail up front and solves it in the cloud with quantum-ready algorithms. Bench and phase sequencing, dig precedence, haulage and processing capacity, grade targets, stockpile strategy, and the capital and operating trade-offs behind them all stay in the model.

Runs cost nothing beyond the credit already on your account. Test the price decks, the delayed capital project, the changed cut-off grade, then take your executives or clients a recommendation you can defend.

Fast schedulers usually get their speed by dropping the constraints that make a plan executable. Cognomine holds drill-and-blast lag, digger separation, slope constraints and haulage in the same solve, so the schedule moves equipment the way a crew would.

How the constraints are held →

Most of the block models that come into Cognomine belong to someone other than the person uploading them. The platform is built for that. Client data stays in Australia, is encrypted in transit and at rest, is isolated to your account, and is deleted when your entitlement ends.

You can hand results to a client read-only with no limit on recipients, and bring up to two additional authorised users in to work on the model alongside you. Nobody you share with has to buy anything. If a client's security team wants to review our controls before you upload, we will support that review.

Read the security and data page →