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-04expert-01 →

Hard #05 sudoku race

27 clues · seed 3005 · 19 models · suite v1

19 AI decision models raced to solve this hard sudoku (27 clues), one digit at a time. Jev 1.14 by TypeSafe came first with 7 mistakes.

0.0s / 83.9s0/19 done

Results

#ModelMistakesCallsCostSolve timeStatus
1Jev 1.14TypeSafe723$0.0051956.4sSolved
2GPT-6 Luna DecisionsOpenAI1227$0.015911sSolved
3Span-02Respan1226$0.00435858sSolved
4Jev 1.13TypeSafeolder1429$0.0068077.2sSolved
5d1 UltraLiquid1426$0.0031348.7sSolved
6Kev 4BJared Palmer1429$0.00365418sSolved
7Clef FlashCloudflare1429$0.00541319sSolved
8Mercury DecideInception1528$0.00347714sSolved
9Decision-1Microsoft1629$0.00385417sSolved
10Clef OmniCloudflare1732$0.032613sSolved
11ClefCloudflare1731$0.011523sSolved
12Solar Decide FlashUpstage1732$0.045154sSolved
13Drex v1.5Nace.AIolder1932$0.00382529sSolved
14Drex v2.0Nace.AI1934$0.00765429sSolved
15d1Liquid2133$0.00206510sSolved
16Decider Flash v1.2Perplexity2134$0.00744155sSolved
17Tev1 4B ExperimentalTogether2437$0.033940sSolved
18Decider v1.2Perplexity2639$0.00323184sSolved
19Classify-1ScaleDown2942$0.051144sSolved