run-0639
capabilitycontaminated — inc-0010, inc-0011, inc-0012
citable URL: https://halobench.com/records/run-0639/ — this address never moves; the anchor /records/#run-0639 keeps resolving
date 2026-08-30 · window 2026-08-29T23:22:03Z → 2026-08-30T02:13:43Z
suite tau2-airline-full26 · config cfg-0180
depth 262,144 tokens · scoring deterministic
Metrics
| metric | value |
|---|---|
| empty_arg_calls | 0 |
| tasks_passed | 19 |
| tasks_total | 24 |
| mean_score | 0.7916666667 |
Guard chain
waiver — Standard full Tau2 arm followed the 5/5 smoke. Capability is N=1 and retains every attempt; no repeated model output was used to improve score.
Energy
metered window: eng-0276
Cited by — computed at build time, never stored
claims clm-0127
model pages qwen38-flash-next