SUDOKU BENCHMARK ยท DEVIOUS

gpt-oss 20B on Devious grid 1

Changed a given digit: 46 of 55 blank cells right, 3 given digits changed. It said it was 100% sure. All of gpt-oss 20B's answers.

RESULTChanged a given digit
CELLS RIGHT46 / 55
SAID100% sure
MODEL TIME57.4 min
COSTlocal, no list price
TOKENS IN / OUT340 / 15,941

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

The end of its reply

<grid>
873642915
591837462
246195783
137256984
924158576
685479321
368912547
419725638
752364819
</grid>
<confidence>100</confidence>

Grid devious-0 of set v1, 26 givens. Puzzle as one line: ..3..2..5.9...746....59.78.....5.89..241..5..6........3..9.........2.6.8.5.3.....