run-0432
capability
citable URL: https://halobench.com/records/run-0432/ — this address never moves; the anchor /records/#run-0432 keeps resolving
date 2026-08-17 · window 2026-08-17T16:54:29Z → 2026-08-17T16:59:20Z
suite tau2-bench-airline@668d3bc · config cfg-0110
depth 32,768 tokens · scoring deterministic
Metrics
| metric | value |
|---|---|
| tasks_passed | 4 |
| tasks_total | 5 |
| mean_score | 0.8 |
turns: total 128 · median 22 · p90 54 · max 54 — machine-written trailer, parsed at build time
Guard chain
guarded by run-0394
Energy
metered window: eng-0167
Per-task trace — the full record, no truncation
| task | reward | turns | seconds |
|---|---|---|---|
| 0 | 1 | 10 | 33 |
| 1 | 0 | 22 | 39 |
| 2 | 1 | 54 | 118 |
| 3 | 1 | 14 | 32 |
| 4 | 1 | 28 | 65 |
reward null = infrastructure error, excluded from every paired computation
Cited by — computed at build time, never stored
claims clm-0082