Span-02

Respan · respan/span-02 · added 20 Oct 2026

Span-02 by Respan ranks #3 of 17 AI decision models overall on DecisionBench: #1 of 17 on Sudoku, #8 of 17 on Rulebook, #9 of 17 on Consistency, #11 of 17 on Odds. View on OpenRouter.

Overall: #3 of 17
Overall score: 75 / 100
Sudoku: #1 of 17
Rulebook: #8 of 17
Consistency: #9 of 17
Odds: #11 of 17

Logical reasoningSudoku#1

Accuracy: 86.2%
Solved: 20/20
Mistakes / puzzle: 8.1

Full Sudoku results →Following written rulesRulebook#8

Accuracy: 95.0%
Confidence gap: 0.029
Cost / 1k tickets: $0.0177

Full Rulebook results →Not getting thrown offConsistency#9

Score: 99.0%
Same if re-asked: 100.0%
Obeyed planted text: 5.0%

Full Consistency results →Weighing riskOdds#11

Score: 37.0%
Value captured: 64.0%
Odds within 2×: 10.0%

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.