FORTRESS (benign) leaderboard
As of November 14, 2026, among the results listed here, Thinking Machines Inkling Small has the highest reported FORTRESS (benign) score, 96.9%, published by Thinking Machines, ahead of Thinking Machines Inkling at 95.9%. 3 models have a FORTRESS (benign) result in our records.
Scores as the model makers published them, not re-run by us; each row links to its announcement.
Best reported score
96.9%
T Thinking Machines Inkling Small
Reported by Thinking Machines, July 30, 2026
3 models ranked · 1 maker
| Rank | Model | Score | Setting | Reported by | Date |
|---|---|---|---|---|---|
| 1 | T Inkling Small | 96.9% | Not published | Thinking Machines | July 30, 2026 (model release) |
| 2 | T Inkling | 95.9% | Not published | Thinking Machines | July 30, 2026 (model release) |
| 3 | T Inkling 2 Small | 79.7% | Not published | Thinking Machines | November 9, 2026 (model release) |
How much to trust these numbers
- Source
- Reported by the model makers themselves. Not re-run by LLMBenchmark or anyone independent.
- Settings
- All ranked results use the same reported setting.
- Spread
- 96.9% to 79.7% across 3 models.
- Freshness
- Newest result published November 9, 2026; oldest July 30, 2026.
What FORTRESS (benign) measures
FORTRESS (benign) tests whether models answer safe requests resembling harmful ones; higher percentages mean fewer unnecessary refusals.