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

← Overall ranking

Drex v1.5

older

Nace.AI · nace-ai/drex-v1.5 · added 20 Oct 2026

Newest version: Drex v2.0. Ranks on this page count every version, older ones included.

Drex v1.5 by Nace.AI, an older version of Drex v2.0, ranks #13 of 19 AI decision models overall on DecisionBench: #12 of 19 on Sudoku, #13 of 19 on Rulebook, #9 of 19 on Consistency, #13 of 19 on Odds. View on OpenRouter.

Overall#13Sudoku#12Rulebook#13Consistency#9Odds#13
Overall
#13 of 19
Overall score
50 / 100
Sudoku
#12 of 19
Rulebook
#13 of 19
Consistency
#9 of 19
Odds
#13 of 19
Logical reasoningSudoku#12
Accuracy
77.6%
Solved
20/20
Mistakes / puzzle
14.5
Full Sudoku results →
Following written rulesRulebook#13
Accuracy
85.0%
Confidence gap
0.097
Cost / 1k tickets
$0.0243
Full Rulebook results →
Not getting thrown offConsistency#9
Score
99.3%
Same if re-asked
100.0%
Obeyed planted text
1.7%
Full Consistency results →
Weighing riskOdds#13
Score
36.7%
Value captured
65.0%
Odds within 2×
8.3%
Full Odds results →

On each benchmark the best model scores 100 and the worst 0. The overall score averages all 4, so every benchmark counts the same. How ranking works.