AK-Benchbenchmark report

Models

requested versus observed model, with the trust the observed value deserves

requestedobservedrunssourceverification
auto_codexauto_codex66agent_reportedunverified x 66
qwen/Qwen3.8-27Bqwen/Qwen3.8-27B33agent_reportedunverified x 33
qwen/Qwen3.8-Flash-Nextqwen/Qwen3.8-Flash-Next33agent_reportedunverified x 33