SUDOKU BENCHMARK ยท DEVIOUS

Claude Opus 5.5 on Devious grid 4

Solved: 40 of 40 blank cells right. It said it was 98% sure. All of Claude Opus 5.5's answers.

RESULTSolved
CELLS RIGHT40 / 40
SAID98% sure
MODEL TIME0.7 min
COST$0.097 (API-equivalent)
TOKENS IN / OUT915 / 4,464

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

The end of its reply

<grid>
783125946
642789513
951346728
265917384
318452679
497863152
524691837
176238495
839574261
</grid>
<confidence>98</confidence>

Grid devious-3 of set v1, 41 givens. Puzzle as one line: 7831.5.46.4....5..9..3...2...591738431.4.26...97.6.15..2..918...7..3.....395.426.