{ "report": "HealthAdminBench Results - Omni Agent Hub (July 2026)", "system_under_test": "Omni Agent Hub, production build", "evaluation_window_utc": { "start": "2026-07-13T18:13:47Z", "end": "2026-07-13T19:30:49Z" }, "benchmark": { "name": "HealthAdminBench v2 (Stanford School of Medicine, Shah Lab)", "license": "Apache-2.0", "tasks": 135, "upstream_commit": "e71a8f4d6923037805b7f51fbbf608d12ea56cf5", "corpus_snapshot_sha256": "b2bafeda3388b0d08c27646ad1950ff89e16dbb05ffe2cb3e1c27eee9fe30b63", "corpus_access": "vendored snapshot; tasks never fetched at evaluation time; per-run snapshot hash recorded in run metadata and identical across all 135 records of this run" }, "build_identification": { "source_revision": "13d371c5ceeb3a428325e49ea455aad624e1a2ae", "container_image_digest": "sha256:5d46c666baee0a038f032bf9daecdb41cdc5972e29b48592f8f1e9827da36468", "freshness": "verified at dispatch: the image tagged with this exact source revision carries this digest; re-verified against the container registry during report preparation" }, "protocol": { "attempts_per_task": 1, "candidate_selection": "none", "pass_semantics": "full credit only (score = 1.0); denominator = sum of the task's check weights", "scoring_input": "benchmark environment recorded end state only; no self-report graded", "unmeasurable_tasks": 0, "isolation": "fresh browser context and freshly reseeded environment state per task" }, "headline": { "full_credit": 91, "tasks": 135, "completion_rate": 0.674, "completion_rate_ci95": [ 0.591, 0.747 ], "mean_task_score": 0.9312 }, "per_task_scores": { "appeals_denials/denial-easy-1.json": 1.0, "appeals_denials/denial-easy-10.json": 1.0, "appeals_denials/denial-easy-11.json": 1.0, "appeals_denials/denial-easy-12.json": 1.0, "appeals_denials/denial-easy-13.json": 1.0, "appeals_denials/denial-easy-14.json": 0.889, "appeals_denials/denial-easy-15.json": 1.0, "appeals_denials/denial-easy-16.json": 1.0, "appeals_denials/denial-easy-17.json": 0.75, "appeals_denials/denial-easy-18.json": 1.0, "appeals_denials/denial-easy-19.json": 1.0, "appeals_denials/denial-easy-2.json": 1.0, "appeals_denials/denial-easy-20.json": 0.444, "appeals_denials/denial-easy-3.json": 1.0, "appeals_denials/denial-easy-4.json": 1.0, "appeals_denials/denial-easy-5.json": 1.0, "appeals_denials/denial-easy-6.json": 1.0, "appeals_denials/denial-easy-7.json": 0.857, "appeals_denials/denial-easy-8.json": 1.0, "appeals_denials/denial-easy-9.json": 0.857, "appeals_denials/denial-hard-1.json": 0.722, "appeals_denials/denial-hard-10.json": 0.9, "appeals_denials/denial-hard-11.json": 0.909, "appeals_denials/denial-hard-12.json": 0.909, "appeals_denials/denial-hard-13.json": 1.0, "appeals_denials/denial-hard-14.json": 0.846, "appeals_denials/denial-hard-15.json": 1.0, "appeals_denials/denial-hard-16.json": 0.929, "appeals_denials/denial-hard-17.json": 0.846, "appeals_denials/denial-hard-18.json": 0.923, "appeals_denials/denial-hard-19.json": 1.0, "appeals_denials/denial-hard-2.json": 0.917, "appeals_denials/denial-hard-20.json": 0.462, "appeals_denials/denial-hard-3.json": 0.9, "appeals_denials/denial-hard-4.json": 0.6, "appeals_denials/denial-hard-5.json": 0.588, "appeals_denials/denial-hard-6.json": 0.444, "appeals_denials/denial-hard-7.json": 0.455, "appeals_denials/denial-hard-8.json": 0.667, "appeals_denials/denial-hard-9.json": 0.5, "appeals_denials/denial-medium-1.json": 0.833, "appeals_denials/denial-medium-10.json": 0.941, "appeals_denials/denial-medium-11.json": 0.778, "appeals_denials/denial-medium-12.json": 0.875, "appeals_denials/denial-medium-13.json": 0.75, "appeals_denials/denial-medium-14.json": 1.0, "appeals_denials/denial-medium-15.json": 1.0, "appeals_denials/denial-medium-16.json": 0.889, "appeals_denials/denial-medium-17.json": 0.889, "appeals_denials/denial-medium-18.json": 1.0, "appeals_denials/denial-medium-19.json": 1.0, "appeals_denials/denial-medium-2.json": 0.895, "appeals_denials/denial-medium-20.json": 1.0, "appeals_denials/denial-medium-3.json": 0.944, "appeals_denials/denial-medium-4.json": 0.944, "appeals_denials/denial-medium-5.json": 0.895, "appeals_denials/denial-medium-6.json": 0.273, "appeals_denials/denial-medium-7.json": 0.3, "appeals_denials/denial-medium-8.json": 1.0, "appeals_denials/denial-medium-9.json": 1.0, "dme/fax-easy-1.json": 1.0, "dme/fax-easy-2.json": 1.0, "dme/fax-easy-3.json": 1.0, "dme/fax-easy-4.json": 1.0, "dme/fax-easy-5.json": 1.0, "dme/fax-hard-1.json": 1.0, "dme/fax-hard-2.json": 1.0, "dme/fax-hard-3.json": 1.0, "dme/fax-hard-4.json": 1.0, "dme/fax-hard-5.json": 0.9, "dme/fax-medium-1.json": 1.0, "dme/fax-medium-2.json": 1.0, "dme/fax-medium-3.json": 1.0, "dme/fax-medium-4.json": 1.0, "dme/fax-medium-5.json": 1.0, "prior_auth/emr-easy-1.json": 1.0, "prior_auth/emr-easy-10.json": 1.0, "prior_auth/emr-easy-11.json": 1.0, "prior_auth/emr-easy-12.json": 1.0, "prior_auth/emr-easy-13.json": 1.0, "prior_auth/emr-easy-14.json": 1.0, "prior_auth/emr-easy-15.json": 1.0, "prior_auth/emr-easy-16.json": 1.0, "prior_auth/emr-easy-17.json": 1.0, "prior_auth/emr-easy-18.json": 1.0, "prior_auth/emr-easy-19.json": 1.0, "prior_auth/emr-easy-2.json": 1.0, "prior_auth/emr-easy-20.json": 1.0, "prior_auth/emr-easy-3.json": 1.0, "prior_auth/emr-easy-4.json": 1.0, "prior_auth/emr-easy-5.json": 1.0, "prior_auth/emr-easy-6.json": 1.0, "prior_auth/emr-easy-7.json": 1.0, "prior_auth/emr-easy-8.json": 1.0, "prior_auth/emr-easy-9.json": 1.0, "prior_auth/emr-hard-1.json": 1.0, "prior_auth/emr-hard-10.json": 1.0, "prior_auth/emr-hard-11.json": 1.0, "prior_auth/emr-hard-12.json": 1.0, "prior_auth/emr-hard-13.json": 1.0, "prior_auth/emr-hard-14.json": 1.0, "prior_auth/emr-hard-15.json": 0.947, "prior_auth/emr-hard-16.json": 1.0, "prior_auth/emr-hard-17.json": 0.864, "prior_auth/emr-hard-18.json": 0.889, "prior_auth/emr-hard-19.json": 1.0, "prior_auth/emr-hard-2.json": 1.0, "prior_auth/emr-hard-20.json": 0.909, "prior_auth/emr-hard-3.json": 1.0, "prior_auth/emr-hard-4.json": 1.0, "prior_auth/emr-hard-5.json": 0.944, "prior_auth/emr-hard-6.json": 0.895, "prior_auth/emr-hard-7.json": 1.0, "prior_auth/emr-hard-8.json": 0.917, "prior_auth/emr-hard-9.json": 1.0, "prior_auth/emr-medium-1.json": 1.0, "prior_auth/emr-medium-10.json": 1.0, "prior_auth/emr-medium-11.json": 1.0, "prior_auth/emr-medium-12.json": 1.0, "prior_auth/emr-medium-13.json": 1.0, "prior_auth/emr-medium-14.json": 1.0, "prior_auth/emr-medium-15.json": 1.0, "prior_auth/emr-medium-16.json": 1.0, "prior_auth/emr-medium-17.json": 1.0, "prior_auth/emr-medium-18.json": 1.0, "prior_auth/emr-medium-19.json": 1.0, "prior_auth/emr-medium-2.json": 1.0, "prior_auth/emr-medium-20.json": 0.926, "prior_auth/emr-medium-3.json": 1.0, "prior_auth/emr-medium-4.json": 1.0, "prior_auth/emr-medium-5.json": 1.0, "prior_auth/emr-medium-6.json": 1.0, "prior_auth/emr-medium-7.json": 1.0, "prior_auth/emr-medium-8.json": 1.0, "prior_auth/emr-medium-9.json": 1.0 } }