GPT-6 Luna

Results

Score

095%: 0 to 9

Grids solved

0 of 42

Blank cells filled correctly

8.6%

Wrong cells / empty cells / changed a given / no grid

3 / 5 / 20 / 14

Broke the rules (complete grid repeating a digit)

3

Near misses (three cells or fewer wrong)

0

Average stated confidence

8%

Average confidence when it had not solved it

8%

Sure but wrong (90% or more, unsolved)

3

Calibration (Brier, lower is better)

0.075

API-equivalent cost per grid

$0.0024

API-equivalent cost per solved grid

n/a

Seconds per grid

24

Tokens per grid (in / out)

56,078 / 1,779

Reasoning tokens per grid

1,234

Retries for an unreadable reply

24

Aborted on our side (not counted)

0

By tier

Cells by the technique they needed

Each blank cell is tagged with the simplest technique that fills it, given everything a solver could fill before it.