SUDOKU BENCHMARK · TRICKY

Claude Opus 5.5 on Tricky grid 3

Solved: 57 of 57 blank cells right. It said it was 99% sure. All of Claude Opus 5.5's answers.

RESULTSolved
CELLS RIGHT57 / 57
SAID99% sure
MODEL TIME2.9 min
COST$0.401 (API-equivalent)
TOKENS IN / OUT907 / 19,667

5 given 5 filled correctly 53 wrong, with the right digit 3 left empty

The end of its reply

<grid>
847936512
925741638
631852749
316498275
589327461
472165983
753689124
268514397
194273856
</grid>

<confidence>99</confidence>

I solved this by hand, testing cases where needed. I then checked that every row, column and 3x3 box contains 1–9 exactly once and that all 24 givens are unchanged.

Grid tricky-2 of set v1, 24 givens. Puzzle as one line: .4...65..9....1..8..1..2.4....4.8....893...........9..75.6...24.68.........2...5.