run-0104
capability
citable URL: https://halobench.com/records/run-0104/ — this address never moves; the anchor /records/#run-0104 keeps resolving
date 2026-08-11 · window 2026-08-11T06:51:22Z → 2026-08-11T07:53:51Z
suite tau2-bench-airline@668d3bc · config cfg-0030
depth 32,768 tokens · scoring deterministic
Metrics
| metric | value |
|---|---|
| tasks_passed | 11 |
| tasks_total | 12 |
| mean_score | 0.9167 |
turns: total 252 · median 24 · p90 26 · max 32 — machine-written trailer, parsed at build time
Guard chain
not a performance run — no guard required
Energy
metered window: eng-0040
Per-task trace — the full record, no truncation
| task | reward | turns | seconds |
|---|---|---|---|
| 0 | 1 | 12 | 97 |
| 1 | 1 | 24 | 194 |
| 2 | 1 | 22 | 192 |
| 3 | 1 | 18 | 102 |
| 4 | 1 | 24 | 159 |
| 5 | 1 | 24 | 195 |
| 6 | 1 | 12 | 86 |
| 8 | 1 | 32 | 226 |
| 9 | 1 | 16 | 155 |
| 10 | 1 | 26 | 717 |
| 11 | 0 | 24 | 196 |
| 12 | 1 | 18 | 206 |
reward null = infrastructure error, excluded from every paired computation
Cited by — computed at build time, never stored
claims clm-0045
model pages qwen35-122b