How to solve the FTO (Face-Turning Octahedron)

KRU

MIND

Speedcubing Tracker

How to solve the FTO (Face-Turning Octahedron)

The FTO is the face-turning octahedron: eight triangular faces that rotate independently. This guide covers the 5-step beginner method, advanced methods, techniques, hardware and the best community resources. The Kruxmind app includes FTO scrambles, a timer and a trainer.

FTO notation

Scrambles use Gottlieb notation (the one in csTimer and Kruxmind): the 8 faces are U, F, R, L, B, D, BR and BL. A letter alone is a 120° clockwise turn of that face; with an apostrophe (') it's counterclockwise.
There are no double turns: each face has only three positions, so two clockwise turns equal one counterclockwise turn.
Bencisco algorithms are read with a corner in front (white on top, green in front) and add suffixes: w = wide turn, s = slice only, o = rotation of the whole puzzle around that face and T = 90° rotation around a corner.

Solving methods

Beginner (5 steps)

Minimal algorithms, no speed. You only need Sledge and Hedge.

White face without corners: pair edges and triangles; build a large trapezoid and a small one and join them, respecting the order of neighboring colors.
Two triples: join each white corner with its 2 triangles and place it without breaking what is done.
The 3 remaining centers: move them using top/side/middle layers (Yau 4×4 style). Colors per the scheme.
Last white corner: build the last white triple; use Sledge/Hedge to reorient.
Last corners and triples: orient and place the triples with Sledge/Hedge; solve the last layer.

FreshCuber Intermediate

Bridge to speed with minimal learning cost. Only 2 algorithms: U-Perm and Triangle-Cycle (plus Sledgehammer).

3 corners of one face (intuitive, or the full face via 2 trapezoids forming the hexagon).
3 opposite corners with Sledgehammer.
Edges with setup moves and Edge-Cycle (U-Perm).
Triangles with Triangle-Cycle.

Bencisco (advanced)

Reduction and blockbuilding, 4 base algorithms. It is the speedsolving method used to set records.

First Center: build one hexagon (large trapezoid plus small one). About 6–8 moves.
First Two Triples (F2T): 2 triples next to the center (Yau edge-building style), 10–15 moves. Mind the triangle orbit; place the 1st triple at the back (next to BR) to insert the 2nd with U or F.
Second Center: only R and top layers. Orient the center (it no longer rotates afterwards). 12 moves or fewer.
Last Two Centers (L2C): the final 2 centers; solving one leaves the other. It is the least efficient step (many wide moves). Advanced: backwards center building.
Last Bottom Triple (LBT): complete the missing triple of the 1st layer with flips, without breaking centers (12–16 moves). Avoid the corner dropping into the keyhole unsolved.
Last Three Triples (L3T): the hardest, in 3 sub-steps. (a) Pair formation: build the 3 triples with flips, avoiding the 2-flip case. (b) Triangle permutation: 1 flip, recognized by the color of the F face. (c) Corner permutation: cycle the 3 corners with CP-CW/CCW.

Nautilus (alternative advanced)

Alternative method by Athefre (2024) in a CFOP/Yau style: more algorithmic, with better lookahead/ergonomics for people coming from NxN cubes. Movecount comparable to Bencisco.

First center (≈ Yau): build the first hexagon.
First triples (≈ Yau): build the triples next to the center.
Last slot (≈ CFOP): solve the final slot of the puzzle body.
Last layer (≈ CFOP): solve the last layer algorithmically.

Techniques, hardware & practice

Advanced techniques

Offset Pair Formation (OPF): in the L3T the triangles don't have to match their corner; it's enough that the last flip permutes them correctly. It works like a last-layer premove and lowers the L3T from ~4 to ~3 flips on average.
OPF rule: when forming your first 2 pairs use the 3 bottom colors at least once, without corner and triangle matching directly. In the final CP the algorithm is the opposite of the layer orientation one (the premove is absorbed).
OPF cost: harder recognition and lots of practice required. (It took Streeter to his 28.42 single.)
Backwards Center Building (in L2C): for the last 2 centers, instead of building the center on the F face, build it on the top face. It's harder (it ends up at BR, where you can't see the pieces) but sometimes saves moves and wide moves on the bad cases.

L3T special cases

1-flip case: you end up with the last triangle forced on top next to the finished pairs and can't close the last pair. Fix: turn U to place the finished pair over the bad triangle, form the pair ignoring the one you almost had, and do an extra flip. Costs 1 extra flip.
2-flip case (always avoid it): happens when the last unpaired triangle falls into its correct slot too early, forcing a pure orientation that needs 2 flips instead of 1. To avoid it, make sure that triangle ends in a different slot after the last flip; there are always 2 options (sledge or hedge), pick the one that does NOT leave it in place.

Recognition

Triangle permutation: look at the color of the F face. If its triangle is up, sledge; if it isn't up (it's paired below), hedge. Almost instant with practice.
Corner permutation: orient the last layer; the direction of that turn is the direction of the 1st move of the correct algorithm. (With OPF, inverted.)

Efficiency per step

First Center: use inspection well; try to plan the whole center before starting.
F2T: build the triples that go in the bottom slots, not the top ones; it eases the transition to the centers.
Second Center: sometimes it's worth leaving the gray edge trapezoid for last, so you don't have to orient the center (saves up to 3 moves).
L2C: it's the least developed step and the most important for fast times. The fewer wide moves, the better; there's still a lot of room for improvement.
LBT: the sledge is more useful than the hedge here (it avoids reorienting the puzzle). If a triangle is already paired with the corner, try to get that side into the R layer after orienting the center (saves a flip).
L3T: avoid the 2-flip case at all costs. With 3 or fewer triangles unsolved, commutators can close everything in one sequence. There are ways to skip the final CP in certain cases.
Ben's philosophy: ~70% of his improvement was solve volume (hundreds), ~30% the method.

Lookahead

Like in NxN: while executing one step, already look at the pieces of the next.
In FTO the critical lookahead is in the L3T pair formation (anticipating which flip leaves you the better case) and in the F2T → centers transition.

Hardware 2026

After years of only the LanLan (poor, needed modding), modern hardware arrived in 2024 and changed everything.
DaYan FTO (Mag + Ball-Core): the GAN of FTOs; smooth, stable, almost no pop; large (~61 mm); top records. For serious speedsolving.
DianSheng Galaxy FTO NEO (Mag + Ball-Core, 2025): small (~56 mm), fast and snappy, pop-resistant, good corner cutting. For speedsolving and small hands.
DianSheng Galaxy FTO v1 (2024): 1st mass-produced magnetic; good but pops; anti-pop kits exist. Budget option.
QiYi FTO (Mag, Tiled): cheap, quiet, smooth. For casual and getting started.
2×2 FTO (DianSheng Galaxy 2×2): the mini, dual of the Skewb. For fun; easy if you already know FTO.
Master / 4×4 FTO (MF8): expert version, Yau method; hardware still rough (pop). Advanced challenge.
Recommendation: serious → DaYan or DianSheng NEO; to try it → QiYi. All (except LanLan) share the color scheme.

Practice and simulators

csTimer: select FTO in the puzzle menu; it gives community random-state scrambles in Gottlieb notation. (No official WCA scrambler yet; it arrives with the 2027 regulations cycle.)
ftosim: web simulator used by the community to practice and generate case diagrams (find it in the Speedsolving FTO thread). Trick: in the console `sim_size = 4; sim_reset();` switches it to Master.
Routine: hundreds of solves. Time yourself, identify the slowest step (usually L2C or L3T) and work on it in isolation.

The FTO in depth

What it is

Octahedron with 8 turnable triangular faces, deep-cut, with no fixed centers (you must memorize the color scheme).
Geometrically it's a Pyraminx superimposed on itself: that's why only 4 colors mix with the other 4. Removing the middle layers it would be a Skewb Diamond.
42 external pieces: 6 corners (4 stickers, 2 orientations), 12 edges (2 stickers, no orientation) and 24 triangles (1 sticker, no orientation, in 2 orbits that never mix).

WCA status

On 24-Jun-2026 the WCA announces that FTO will be an official event.
Competable from 2-Jan-2027 (not before). Format: Average of 5.
Specific regulations still not finalized (January 2027 regulations cycle; community proposals until 25-Jul-2026). No official WCA scrambler yet.
It's the first new category since Skewb (2014).
The same announcement retires the Rubik's Clock after the 2027 Worlds (18-Jul-2027), which sparked controversy.

History

1980s: the idea starts soon after the success of Rubik's cube. Ernő Rubik himself showed interest and patented a version in Hungary (Oct-1980); international patent in 1981.
1982: patents by Hewlett (USA) and Rohrbach (Germany); neither became a product.
1997: Xie Zongliang (Taiwan) patents it; ~1,000 units by 2008.
2001–2003: David Pitcher develops a working mechanism; patents in 2003 but doesn't pay fees, so it falls into the public domain (key to others manufacturing it).
2008–2009: first commercial productions; cult piece for collectors.
1-Apr-2018: Ben Streeter starts the Bencisco method. On 23-Apr-2018 he achieves the first documented sub-1-minute solve (previous record: Michael Gottlieb, 1:04.86).
2020–21 (COVID): interest explodes — Bencisco tutorial on YouTube, leaderboard and FTO Fan Club Discord. Streeter creates the RexTO (printable extensions over a Rex Cube).
2024: modern hardware (DianSheng, DaYan) makes the RexTO obsolete and boosts adoption.
2026: the WCA makes it official.

Records

No official WCA records yet (the event isn't competable until 2027); these are by community consensus.
The world's best are around sub-12 s singles and ~14–15 s averages (2024–2026). Historical reference: in 2021 you already saw ~18.58 single / ~19.80 ao5.
Live rankings on cubingcontests.com and the Speedsolving wiki (they change often).
Master FTO (4×4): first sub-2 achieved by Ledger (~1:46 single).

Community and trivia

Active community on the FTO Fan Club Discord and the FTO threads of Speedsolving.com.
FTO and Skewb are duals (the 2×2 FTO feels like a Skewb).
The method is called Bencisco as a play on Ben Streeter's name.
Recurring joke: all that's missing is a speed-solvable icosahedron to complete the Platonic solids.
2026 drama: FTO is celebrated as an event, but the Clock community protests its simultaneous removal.

Practising the FTO on Kruxmind

Timer: pick FTO and time yourself with an Ao5, the official format. Scrambles are random-state: a position of the FTO is picked at random and the way to reach it is computed, so every position is equally likely.
Scramble generator: numbered batches of 5, 12 or 25 FTO scrambles, to practise a round with friends or run a small competition.
Virtual FTO: solve an FTO in 3D with the keyboard or by tapping the screen, without the puzzle at hand.
Algorithms and trainer: the app has the Bencisco triggers, FreshCuber's intermediate method, the full L3T (10 pair formation, 5 top layer and 3 last triangles), the 18 TCP cases and the Nautilus sets (L3C, plus L6X and PL6X with 35 cases each).

Frequently asked questions

Is the FTO an official WCA event?

Yes. The WCA announced it on 24 June 2026 and it can be competed in from 2 January 2027, with an average of 5 (Ao5). It's the first new event since Skewb in 2014.

Which method should I start with?

With the 5-step beginner method: you only need two algorithms, Sledge and Hedge. When you want faster times, move to Bencisco, the method behind the records; Nautilus is the alternative if you come from CFOP or Yau.

Which FTO should I buy?

To compete, the DaYan FTO or the DianSheng Galaxy FTO NEO; to try it without spending much, the QiYi FTO. They all share the same colour scheme.

Resources & links

← Home · Dictionary · Guides · Novedades · FAQ · Smart cubes · Solving methods · Scientific basis · About