Model library · GPT-4.1 mini · 2026-10-28
Airline simulations
Tests instruction following over a long booking policy.
| Task | Trial | Result | Ended | Messages | Cost | |
|---|---|---|---|---|---|---|
| 1 | 0 | fail | user_stop | 24 | $0.018 | Transcript → |
| 12 | 0 | fail | user_stop | 36 | $0.037 | Transcript → |
| 14 | 0 | fail | user_stop | 23 | $0.024 | Transcript → |
| 17 | 0 | fail | user_stop | 91 | $0.156 | Transcript → |
| 18 | 0 | fail | user_stop | 33 | $0.029 | Transcript → |
| 19 | 0 | fail | user_stop | 32 | $0.030 | Transcript → |
| 21 | 0 | fail | user_stop | 46 | $0.058 | Transcript → |
| 22 | 0 | fail | user_stop | 23 | $0.018 | Transcript → |
| 23 | 0 | fail | user_stop | 28 | $0.027 | Transcript → |
| 26 | 0 | fail | user_stop | 14 | $0.010 | Transcript → |
| 27 | 0 | fail | user_stop | 26 | $0.016 | Transcript → |
| 29 | 0 | fail | user_stop | 24 | $0.023 | Transcript → |
| 33 | 0 | fail | user_stop | 34 | $0.035 | Transcript → |
| 35 | 0 | fail | user_stop | 23 | $0.024 | Transcript → |
| 36 | 0 | fail | user_stop | 14 | $0.010 | Transcript → |
| 37 | 0 | fail | user_stop | 17 | $0.012 | Transcript → |
| 38 | 0 | fail | user_stop | 22 | $0.019 | Transcript → |
| 39 | 0 | fail | user_stop | 20 | $0.011 | Transcript → |
| 41 | 0 | fail | user_stop | 21 | $0.013 | Transcript → |
| 42 | 0 | fail | user_stop | 36 | $0.036 | Transcript → |
| 43 | 0 | fail | user_stop | 27 | $0.022 | Transcript → |
| 44 | 0 | fail | user_stop | 73 | $0.098 | Transcript → |
| 45 | 0 | fail | user_stop | 22 | $0.018 | Transcript → |
| 47 | 0 | fail | user_stop | 14 | $0.009 | Transcript → |
| 5 | 0 | fail | user_stop | 27 | $0.023 | Transcript → |
| 7 | 0 | fail | user_stop | 37 | $0.033 | Transcript → |
| 8 | 0 | fail | user_stop | 24 | $0.022 | Transcript → |
| 9 | 0 | fail | user_stop | 19 | $0.013 | Transcript → |