Skip to content
decisionbench
OverallSudokuRulebookConsistencyOdds

Independent benchmark. Not affiliated with or endorsed by OpenRouter or any model provider. Company and model names are trademarks of their respective owners.

How ranking worksPrivacyTermsContact

© 2026 editid0

← All races
← hard-03hard-05 →

Hard #04 sudoku race

27 clues · seed 3004 · 19 models · suite v1

19 AI decision models raced to solve this hard sudoku (27 clues), one digit at a time. Span-02 by Respan came first with 7 mistakes.

0.0s / 89.4s0/19 done

Results

#ModelMistakesCallsCostSolve timeStatus
1Span-02Respan723$0.00362051sSolved
2Jev 1.13TypeSafeolder1127$0.0062578.7sSolved
3d1 UltraLiquid1225$0.0029438.9sSolved
4Decision-1Microsoft1328$0.00358016sSolved
5Jev 1.14TypeSafe1427$0.0058458.3sSolved
6Clef FlashCloudflare1429$0.00580322sSolved
7Decider Flash v1.2Perplexity1429$0.00576646sSolved
8d1Liquid1529$0.0017059.5sSolved
9Drex v1.5Nace.AIolder1530$0.00360819sSolved
10Drex v2.0Nace.AI1529$0.00656321sSolved
11Clef OmniCloudflare1731$0.031814sSolved
12ClefCloudflare1731$0.010524sSolved
13Tev1 4B ExperimentalTogether1732$0.030543sSolved
14GPT-6 Luna DecisionsOpenAI2237$0.023915sSolved
15Decider v1.2Perplexity2236$0.00295541sSolved
16Solar Decide FlashUpstage2335$0.052257sSolved
17Mercury DecideInception2536$0.00482824sSolved
18Kev 4BJared Palmer2739$0.00538889sSolved
19Classify-1ScaleDown2938$0.044532sSolved