run-0345
capability
citable URL: https://halobench.com/records/run-0345/ — this address never moves; the anchor /records/#run-0345 keeps resolving
date 2026-08-16 · window 2026-08-16T12:58:58Z → 2026-08-16T14:08:51Z
suite tau2-bench-airline@668d3bc · config cfg-0095
depth 32,768 tokens · scoring deterministic
Metrics
| metric | value |
|---|---|
| tasks_passed | 23 |
| tasks_total | 26 |
| mean_score | 0.8846 |
turns: total 722 · median 26 · p90 38 · max 56 — machine-written trailer, parsed at build time
Guard chain
guarded by run-0338
Energy
metered window: eng-0134
Per-task trace — the full record, no truncation
| task | reward | turns | seconds |
|---|---|---|---|
| 0 | 1 | 12 | 62 |
| 1 | 1 | 28 | 78 |
| 2 | 1 | 22 | 105 |
| 3 | 1 | 16 | 46 |
| 4 | 1 | 20 | 101 |
| 5 | 1 | 22 | 91 |
| 6 | 1 | 24 | 83 |
| 7 | 0 | 34 | 156 |
| 8 | 1 | 26 | 83 |
| 9 | 1 | 20 | 79 |
| 10 | 1 | 28 | 202 |
| 11 | 1 | 24 | 83 |
| 12 | 1 | 26 | 104 |
| 13 | 1 | 36 | 89 |
| 14 | 0 | 32 | 295 |
| 15 | 1 | 30 | 232 |
| 16 | 1 | 22 | 111 |
| 17 | 1 | 30 | 184 |
| 18 | 1 | 34 | 101 |
| 19 | 1 | 22 | 76 |
| 20 | 1 | 26 | 110 |
| 21 | 1 | 38 | 223 |
| 22 | 1 | 34 | 160 |
| 23 | 0 | 38 | 332 |
| 24 | 1 | 56 | 212 |
| 25 | 1 | 22 | 101 |
reward null = infrastructure error, excluded from every paired computation