GPT
Trajectory
Concept trajectory
No baseline to compare against
This is the family's first analyzed generation. Once a second member is benched and analyzed, this section will surface the per-concept delta (resolved / persisting / regressed / new).
Members
| Model | Generation | Pass@N | Avg cost / task | Runs | Last run |
|---|---|---|---|---|---|
| 4 | — | — | 0 | — | |
| 5 | — | — | 0 | — | |
| 5 | 79.4% | $0.13 | 3 | 5mo ago | |
| 5 | 69.3% | $0.02 | 3 | 5mo ago | |
| 5 | — | — | 0 | — | |
| 5 | 89.8% | $0.06 | 3 | 1mo ago | |
| 5 | 84.3% | $0.04 | 3 | 1mo ago | |
| 5 | 77.3% | $0.004 | 3 | 1mo ago | |
| 6 | — | — | 0 | — | |
| 6 | 87.9% | $0.02 | 3 | 24d ago | |
| 6 | 80.0% | $0.002 | 3 | 24d ago | |
| 6 | 93.8% | $0.02 | 3 | 18d ago |