Decision-1
Microsoft · microsoft/microsoft-decision-1 · added 20 Oct 2026
Decision-1 by Microsoft ranks #3 of 12 AI decision models overall on DecisionBench: #7 of 12 on Sudoku, #2 of 12 on Triage, #5 of 12 on Robustness. View on OpenRouter.
- Overall
- #3 of 12
- Average rank
- 4.7
- Sudoku
- #7 of 12
- Triage
- #2 of 12
- Robustness
- #5 of 12
Sudoku Logical reasoning
Leaderboard →- Rank
- #7 of 12
- Accuracy
- 79.6%
- Solved
- 20/20
- Mistakes / puzzle
- 12.9
- Solve time / puzzle
- 15s
- Cost / puzzle
- $0.003344
By difficulty
| Difficulty | Rank | Accuracy | Solved | Mistakes | Solve time | Cost |
|---|---|---|---|---|---|---|
| Easy | #4 | 90.3% | 5/5 | 4.4 | 9.4s | $0.001772 |
| Medium | #7 | 82.2% | 5/5 | 10.6 | 14s | $0.003006 |
| Hard | #8 | 76.1% | 5/5 | 17.0 | 18s | $0.004085 |
| Expert | #7 | 74.5% | 5/5 | 19.6 | 21s | $0.004512 |
Every race (20)
| Puzzle | Mistakes | Calls | Solve time | Cost | Status |
|---|---|---|---|---|---|
| easy-01 | 7 | 19 | 13s | $0.002091 | Solved |
| easy-02 | 3 | 16 | 7.7s | $0.001616 | Solved |
| easy-03 | 3 | 16 | 9.1s | $0.001617 | Solved |
| easy-04 | 5 | 18 | 9.0s | $0.001960 | Solved |
| easy-05 | 4 | 16 | 7.9s | $0.001577 | Solved |
| medium-01 | 9 | 23 | 13s | $0.002719 | Solved |
| medium-02 | 15 | 27 | 16s | $0.003445 | Solved |
| medium-03 | 15 | 29 | 16s | $0.003615 | Solved |
| medium-04 | 7 | 22 | 12s | $0.002654 | Solved |
| medium-05 | 7 | 22 | 13s | $0.002599 | Solved |
| hard-01 | 16 | 32 | 19s | $0.004370 | Solved |
| hard-02 | 21 | 34 | 20s | $0.004596 | Solved |
| hard-03 | 19 | 32 | 18s | $0.004028 | Solved |
| hard-04 | 13 | 28 | 16s | $0.003580 | Solved |
| hard-05 | 16 | 29 | 17s | $0.003854 | Solved |
| expert-01 | 21 | 33 | 19s | $0.004389 | Solved |
| expert-02 | 17 | 31 | 18s | $0.004046 | Solved |
| expert-03 | 23 | 37 | 23s | $0.005141 | Solved |
| expert-04 | 12 | 29 | 17s | $0.003692 | Solved |
| expert-05 | 25 | 38 | 25s | $0.005291 | Solved |
Triage Following written rules
Leaderboard →- Rank
- #2 of 12
- Accuracy
- 100.0%
- Team
- 100.0%
- Refund
- 100.0%
- Urgency
- 100.0%
- Confidence gap
- 0.076
60 tickets · 0.0% of mistakes made while at least 90% sure · $0.0260 per 1,000 tickets
Robustness Consistency under pressure
Leaderboard →- Rank
- #5 of 12
- Score
- 99.3%
- Same if re-asked
- 100.0%
- Changed by reorder
- 0.0%
- In a list of 40
- 98% / 100%
- Obeyed planted text
- 1.7%
20 tickets × 6 checks · changed by rename 0.0% · accuracy with planted text 98.3%