Research records

Runs

A run is one complete execution of the Arena harness: a fixed set of configurations attempting the pinned task pack in isolated sandboxes. Each record keeps blocked, rejected, interrupted, and incomplete states visible. A run still marked running after 24 hours is flagged as likely interrupted.

Run index

newest first
Experimental run records; blocked and rejected outcomes remain visible.
RunStatusTrialsAcceptedCreated
ea17d555accepted484827 Jul 2026, 00:14 UTC
f8a161d8expired0027 Jul 2026, 00:11 UTC
5aeb58d4expired0027 Jul 2026, 00:05 UTC
be0d3ee0cancelled2226 Jul 2026, 20:39 UTC
c55f0f7finconclusive968426 Jul 2026, 20:38 UTC
54385280inconclusive969026 Jul 2026, 20:12 UTC
3eb89892inconclusive241726 Jul 2026, 19:56 UTC
23cad676accepted4312 Jul 2026, 18:16 UTC
90a4a241accepted12812 Jul 2026, 18:10 UTC
df7df63daccepted12612 Jul 2026, 18:10 UTC
4a0b7258accepted12512 Jul 2026, 18:09 UTC