Lab 001 · Benchmark

RubiksBench

Can a language model solve a Rubik's Cube in one reply? Each model sees a scrambled cube as text and must write every move at once. No tools and no second try. Our engine plays the moves and checks the cube.

Models
26
Leader
GPT-6 Astra
Updated
2026-09-24
Scramble
#ModelEffortCubesSolvedMoves$ / cube
01GPT-6 Astraxhighxhigh10/1074 moves74$0.93
02GPT-5.6 Solxhighxhigh7/10101 moves101$0.64
03Claude Opus 5.5lowlow5/1072 moves72$0.42
04Claude Opus 5lowlow4/1031 moves31$3.99
05GPT-6 Solhighhigh2/1054 moves54$0.23
06GPT-6 Lunahighhigh0/10—<$0.01
07Gemini 3.1 Pro Previewlowlow0/10—<$0.01
08Gemini 3.8 Flashlowlow0/10—$0.10
09Claude Sonnet 5lowlow0/10—$0.17
10DeepSeek V4.1 Flashhighhigh0/10—$0.19
11Kimi K3lowlow0/10—$0.28
Provisional · fewer than 10 cubes · not a final score
··GLM 5.3 FlashXlowlow2/972 moves72$0.09
··Claude Fable 5.1maxmax1/522 moves22$7.61
··DeepSeek V4 Flash 0731lowlow0/9—$0.02
··DeepSeek V4 Pro 0813lowlow0/8—$0.06
··Qwen3.8 27Bxhighxhigh0/8—$0.11
··GLM 5.3 Flashlowlow0/6—$0.14
··Qwen3.8 Flashlowlow0/4—$0.13

solvedwrong answerno answer (thought until its budget ran out)

Across scramble lengths · best setting per model

* fewer than 10 cubes. A cube scrambled with 15 or more moves is as mixed as a cube can get, so longer scrambles add no difficulty.

Method
Input
All 54 stickers of the cube, laid out as a flat map with each face labelled.
Prompt
Show the prompt for cube 1 at 6 moves

This is the whole prompt, sent as the only message. There is no system message. Only the cube changes between cubes. Each model page shows the prompt for every cube.

Below is a scrambled Rubik's Cube, shown unfolded with all 6 faces visible. Each face is a 3x3 grid of stickers. Each letter names a sticker's color, using the face it belongs on when the cube is solved: U=White (Up), R=Red (Right), F=Green (Front), D=Yellow (Down), L=Orange (Left), B=Blue (Back).

The diagram is laid out as follows:
- The first 3 lines (indented) are the U (Up) face.
- The next 3 lines each hold four space-separated 3-sticker groups. Left to right, these groups are the L (Left), F (Front), R (Right) and B (Back) faces.
- The last 3 lines (indented) are the D (Down) face.

Each face's rows and columns are oriented as they appear once the cube is folded up:
- U: its first line is the row touching B, and its last line is the row touching F.
- D: its first line is the row touching F, and its last line is the row touching B.
- L, F, R and B: the first line of each is the row touching U.
- Reading a line from left to right: U, F and D run from L to R; R runs from F to B; L runs from B to F; and B runs from R to L.

      U U F
      B U F
      L F R
B D F U R F D L L D L R
L L F U F B U R R D B D
R R F U U D B B B U F D
      R R L
      B D L
      B D L

Give the sequence of moves that returns this cube to a fully solved state (all 9 stickers on each face matching).
Respond with only the move sequence in standard notation: a space-separated list of face letters (U, D, L, R, F, or B), each optionally followed by ' for a counterclockwise turn or 2 for a double turn. Do not copy any example — work out the actual sequence for the cube shown above.
Do not include any explanation, commentary, or code.
Output
One list of moves in standard notation. Nothing else is read.
Check
A cube counts as solved only when every sticker is back in place.
Cubes
Every model gets the same 10 cubes at each scramble length, in the same order.
Effort
Many models let you set how long they think. The board shows each model at its best measured setting. Tick "every effort setting" to see the rest.
Moves
Average length of the correct answers. A 6-move scramble can be solved in about 7 moves.