run-0674
capability
citable URL: https://halobench.com/records/run-0674/ — this address never moves; the anchor /records/#run-0674 keeps resolving
date 2026-10-03 · window 2026-10-03T13:42:22Z → 2026-10-03T14:23:52Z
suite tau2-airline-full26 · config cfg-0188
depth 200,192 tokens · scoring deterministic
Metrics
| metric | value |
|---|---|
| tasks_passed | 22 |
| tasks_total | 26 |
Guard chain
waiver — Formal 4/4 needle guard at depth 8000 not run; pre-flight was a one-task tau2 airline smoke (task 0, reward 1.0) and a 60-case tool-call screen on the same build (59/60). Disclosed deviation.
Energy
metered window: eng-0281
Cited by — computed at build time, never stored
claims clm-0136
model pages qwen38-flash-next