Drex v2.0

Nace.AI · nace-ai/drex-v2.0 · added 20 Oct 2026

Older version: Drex v1.5

Drex v2.0 by Nace.AI ranks #5 of 17 AI decision models overall on DecisionBench: #6 of 17 on Sudoku, #6 of 17 on Rulebook, #3 of 17 on Consistency, #13 of 17 on Odds. View on OpenRouter.

Overall: #5 of 17
Overall score: 67 / 100
Sudoku: #6 of 17
Rulebook: #6 of 17
Consistency: #3 of 17
Odds: #13 of 17

Logical reasoningSudoku#6

Accuracy: 81.5%
Solved: 20/20
Mistakes / puzzle: 11.4

Full Sudoku results →Following written rulesRulebook#6

Accuracy: 96.1%
Confidence gap: 0.030
Cost / 1k tickets: $0.0364

Full Rulebook results →Not getting thrown offConsistency#3

Score: 100.0%
Same if re-asked: 100.0%
Obeyed planted text: 0.0%

Full Consistency results →Weighing riskOdds#13

Score: 34.8%
Value captured: 69.7%
Odds within 2×: 0.0%

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.