Home › Evidence › Records › run-0372

run-0372

capability
citable URL: https://halobench.com/records/run-0372/ — this address never moves; the anchor /records/#run-0372 keeps resolving
date 2026-08-17 · window 2026-08-17T06:37:19Z → 2026-08-17T07:02:15Z
suite deep-thought-posttrain-fullbench · config cfg-0101
depth 8,192 tokens · scoring deterministic

Metrics

metricvalue
tasks_passed5
tasks_total14
mean_score0.3571

turns: total 465 · median 8 · p90 36 · max 201 — machine-written trailer, parsed at build time

Guard chain

not a performance run — no guard required

Energy

metered window: eng-0146

Per-task trace — the full record, no truncation

taskrewardturnsseconds
00201119
11834
2null (infrastructure)0
301613
411650
511612
6null (infrastructure)0
704024
80169
913636
1012015
11null (infrastructure)0
12null (infrastructure)0
13null (infrastructure)0
14null (infrastructure)0
150833
1601237
17null (infrastructure)0
18null (infrastructure)0
19null (infrastructure)0
20null (infrastructure)0
2103226
22null (infrastructure)0
2302216
2402218
25null (infrastructure)0

reward null = infrastructure error, excluded from every paired computation

Cited by — computed at build time, never stored

claims clm-0078
model pages deep-thought-posttrain
candidate gate history deep-thought-posttrain