SUDOKU BENCHMARK · REASONING MEDIUM

Claude Fable 5.1

Anthropic · released 2026-09-01 · route claude-fable-5-1 · Model profile · Chess results

Results

Score
9395%: 83 to 100
Grids solved
39 of 42
Blank cells filled correctly
97.6%
Wrong cells / empty cells / changed a given / no grid
2 / 0 / 1 / 0
Broke the rules (complete grid repeating a digit)
2
Near misses (three cells or fewer wrong)
2
Average stated confidence
97%
Average confidence when it had not solved it
68%
Sure but wrong (90% or more, unsolved)
2
Calibration (Brier, lower is better)
0.047
API-equivalent cost per grid
$0.894
API-equivalent cost per solved grid
$0.963
Seconds per grid
173
Tokens per grid (in / out)
2,235 / 16,999
Reasoning tokens per grid
15,035
Retries for an unreadable reply
1
Aborted on our side (not counted)
0

By tier

TierNeedsSolvedOutput tokensSeconds
CozySingles only6 of 65,12856
GentleScanning6 of 610,788109
SteadyLocked candidates6 of 616,626170
TrickyHidden pairs and triples4 of 611,143122
DeviousWings and fish5 of 69,18893
ExpertDeep wings and fish6 of 623,410250
BeyondNo human technique finishes it6 of 642,711413

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.

TechniqueCellsFilled correctly
Last free cell43299.3%
Scanning67499.4%
Locked candidates274100.0%
Hidden pairs and triples25882.6%
Wings and fish16598.8%
Deep wings and fish181100.0%
Search249100.0%

Every answer

Cozy #1Solved 50 of 50 cells100% sure · 6,935 tokens · $0.365Cozy #2Solved 48 of 48 cells99% sure · 4,617 tokens · $0.249Cozy #3Solved 49 of 49 cells99% sure · 5,501 tokens · $0.293Cozy #4Solved 50 of 50 cells99% sure · 4,992 tokens · $0.268Cozy #5Solved 49 of 49 cells99% sure · 5,080 tokens · $0.272Cozy #6Solved 39 of 39 cells99% sure · 3,644 tokens · $0.200Gentle #1Solved 56 of 56 cells99% sure · 12,739 tokens · $0.655Gentle #2Solved 51 of 51 cells99% sure · 14,829 tokens · $0.760Gentle #3Solved 52 of 52 cells99% sure · 7,537 tokens · $0.395Gentle #4Solved 51 of 51 cells99% sure · 7,561 tokens · $0.396Gentle #5Solved 56 of 56 cells99% sure · 12,705 tokens · $0.654Gentle #6Solved 52 of 52 cells99% sure · 9,360 tokens · $0.486Steady #1Solved 53 of 53 cells99% sure · 11,281 tokens · $0.582Steady #2Solved 53 of 53 cells99% sure · 9,998 tokens · $0.518Steady #3Solved 49 of 49 cells99% sure · 14,410 tokens · $0.739Steady #4Solved 54 of 54 cells98% sure · 27,307 tokens · $1.38Steady #5Solved 53 of 53 cells99% sure · 9,572 tokens · $0.497Steady #6Solved 54 of 54 cells99% sure · 27,190 tokens · $1.38Tricky #1Solved 55 of 55 cells100% sure · 15,148 tokens · $0.775Tricky #2Solved 57 of 57 cells99% sure · 10,700 tokens · $0.553Tricky #3Solved 57 of 57 cells99% sure · 15,031 tokens · $0.770Tricky #4Solved 54 of 54 cells100% sure · 12,679 tokens · $0.652Tricky #5Wrong cells 53 of 55 cells99% sure · 12,423 tokens · $0.639Tricky #6Changed a given digit 6 of 56 cells5% sure · 879 tokens · $0.062Devious #1Solved 55 of 55 cells99% sure · 10,330 tokens · $0.535Devious #2Solved 55 of 55 cells99% sure · 13,552 tokens · $0.696Devious #3Solved 49 of 49 cells99% sure · 8,315 tokens · $0.434Devious #4Wrong cells 38 of 40 cells99% sure · 5,441 tokens · $0.290Devious #5Solved 47 of 47 cells99% sure · 7,648 tokens · $0.401Devious #6Solved 54 of 54 cells99% sure · 9,839 tokens · $0.510Expert #1Solved 58 of 58 cells97% sure · 28,310 tokens · $1.43Expert #2Solved 53 of 53 cells99% sure · 10,177 tokens · $0.527Expert #3Solved 54 of 54 cells97% sure · 43,571 tokens · $2.20Expert #4Solved 57 of 57 cells99% sure · 17,430 tokens · $0.890Expert #5Solved 58 of 58 cells99% sure · 19,958 tokens · $1.02Expert #6Solved 54 of 54 cells99% sure · 21,017 tokens · $1.07Beyond #1Solved 60 of 60 cells99% sure · 78,775 tokens · $5.05Beyond #2Solved 57 of 57 cells99% sure · 23,019 tokens · $1.17Beyond #3Solved 56 of 56 cells99% sure · 51,505 tokens · $2.59Beyond #4Solved 57 of 57 cells99% sure · 18,866 tokens · $0.962Beyond #5Solved 59 of 59 cells97% sure · 35,274 tokens · $1.78Beyond #6Solved 57 of 57 cells99% sure · 48,825 tokens · $2.46