Tev1 4B Experimental
Together · togethercomputer/tev1-4b-experimental · added 20 Oct 2026
Tev1 4B Experimental by Together ranks #14 of 17 AI decision models on the DecisionBench Sudoku benchmark (logical reasoning), with accuracy 75.4%. View on OpenRouter.
- Sudoku rank
- #14 of 17
- Accuracy
- 75.4%
- Solved
- 20/20
- Mistakes / puzzle
- 16.4
- Solve time / puzzle
- 36s
- Cost / puzzle
- $0.0261
By difficulty
| Difficulty | Rank | Accuracy | Solved | Mistakes | Solve time | Cost |
|---|---|---|---|---|---|---|
| Easy | #19 | 83.0% | 5/5 | 8.4 | 20s | $0.0138 |
| Medium | #15 | 78.0% | 5/5 | 13.8 | 33s | $0.0220 |
| Hard | #15 | 73.8% | 5/5 | 19.2 | 48s | $0.0301 |
| Expert | #15 | 70.3% | 5/5 | 24.2 | 44s | $0.0387 |
Every race
| Puzzle | Mistakes | Calls | Solve time | Cost | Status |
|---|---|---|---|---|---|
| easy-01 | 6 | 20 | 19s | $0.0123 | Solved |
| easy-02 | 10 | 22 | 22s | $0.0153 | Solved |
| easy-03 | 5 | 18 | 16s | $0.0109 | Solved |
| easy-04 | 10 | 20 | 20s | $0.0140 | Solved |
| easy-05 | 11 | 23 | 23s | $0.0161 | Solved |
| medium-01 | 15 | 29 | 31s | $0.0258 | Solved |
| medium-02 | 14 | 27 | 31s | $0.0221 | Solved |
| medium-03 | 13 | 26 | 27s | $0.0185 | Solved |
| medium-04 | 9 | 24 | 33s | $0.0180 | Solved |
| medium-05 | 18 | 30 | 44s | $0.0256 | Solved |
| hard-01 | 20 | 35 | 62s | $0.0333 | Solved |
| hard-02 | 16 | 29 | 50s | $0.0248 | Solved |
| hard-03 | 19 | 33 | 45s | $0.0278 | Solved |
| hard-04 | 17 | 32 | 43s | $0.0305 | Solved |
| hard-05 | 24 | 37 | 40s | $0.0339 | Solved |
| expert-01 | 30 | 43 | 50s | $0.0451 | Solved |
| expert-02 | 27 | 44 | 49s | $0.0450 | Solved |
| expert-03 | 15 | 32 | 37s | $0.0293 | Solved |
| expert-04 | 20 | 35 | 39s | $0.0327 | Solved |
| expert-05 | 29 | 41 | 46s | $0.0416 | Solved |
Tev1 4B Experimental across all benchmarks · How Sudoku works