Every event we run

An octahedron that turns about its eight faces. Newly added to the WCA programme and not contested until 2027.

Eight layers. Past this size the solve is less about algorithms and more about keeping a system for hundreds of pieces.

Nine layers, and solves that run well past five minutes. A big-cube specialist's event.

Ten layers. The pieces stop being individually memorable and the method has to carry the whole solve.

Eleven layers, with an odd centre to orient. Among the largest mass-produced cubes anyone competes on.

Twelve layers, and a solve measured in tens of minutes. The largest cube we would run a round for.

A 3x3x3 in one colour, where the pieces are different sizes instead. You solve by shape rather than by colour.

A 3x3x3 cut into a tetrahedron. It shape-shifts as you turn it, so the first job is getting the shape back.

A 3x3x3 rotated forty-five degrees inside its own shell. Familiar to solve, unfamiliar to look at.

A 3x3x3 cut on a slant, so the layers look uneven. The solve is the same; reading the pieces is not.

A 3x3x3 cut along a diagonal axis. It shape-shifts with every turn and rarely looks like a cube mid-solve.

A 3x3x3 in a single colour with every piece a different shape. Solved entirely by fit, with no colour to help.

A cube that turns only about its corners, moving edges alone. Easy to scramble, surprisingly tidy to solve.

Corner turns that carry the edges with them. A short algorithm set and a very approachable first non-cube puzzle.

Four corners and six centres, and nothing else. The gentlest puzzle on this list and a good first competition event.

A cube that turns about its edges rather than its faces, which lets the corners jumble into positions no cube allows.

Turning one face turns the gears between them. Every move is bigger than it looks, and the solve is mostly one long sequence.

The gear mechanism in a sphere. Same idea, no corners to orient yourself by.

A 3x3x3 whose edges and centres can swap places. The extra freedom needs its own small set of algorithms.

A helicopter cube with curved cuts, which lets a face jumble part way and lock the puzzle into a new shape.

A 3x3x3 whose pieces are cut to different heights, so the puzzle stretches and shrinks as you turn it.

A megaminx with corners only — the 2x2x2 of dodecahedra. Fast, and a natural way into the minx family.

A megaminx with two extra layers per face. Hundreds of pieces and solves measured in minutes.

The next step up again: four layers per face, and a solve that is a serious sitting rather than an attempt.

Five layers per face and close to a thousand stickers. An hours-long solve that few people attempt twice.

A kilominx with an extra layer, so edges appear alongside the corners. The step between kilominx and megaminx.

A pyraminx with an extra layer, adding centres and edges to the familiar tips.

A pyraminx reduced to its four corner tips. Solved in a handful of moves, which is exactly the appeal.

Four layers of tetrahedron. The tips and centres are familiar; everything between them is not.

Five layers of tetrahedron, and the largest pyraminx anyone competes on.

A skewb with an extra cut per axis, which adds edges and doubles the work.

The skewb mechanism cut into a dodecahedron. It looks like a megaminx and solves like nothing of the sort.

Square-1 with the top and bottom layers cut into equal wedges. No shape-restoring shortcut survives.

A third layer added to the Square-1 family. Shape-shifting on two planes at once.

The 2x2x2 solved blindfolded. The shortest memorisation in the sport, and often somebody's first blindfolded event.

The 6x6x6 blindfolded. Beyond the WCA programme, and a memorisation load few people carry.

The 7x7x7 blindfolded, and about as far as this discipline has been taken.

A megaminx memorised and solved blindfolded. Twelve faces with nothing to look at.

A pyraminx blindfolded. Short enough to be a genuine entry point into blindfolded solving.