{
  "schema": "anvil-serving.benchmark-public-summary/v1",
  "campaign": "2026-10-16-quasar-swift15",
  "decision": "no-promotion",
  "promoted": false,
  "recommendation": "needs_more_data",
  "human_gate_required": true,
  "tools": {
    "quasar": "16/18",
    "swift15": "18/18"
  },
  "coding": {
    "quasar_list_pair": "119/123; 20/21",
    "quasar_tuple_pair_sensitivity": "123/123; 21/21",
    "swift15_both_readings": "120/123; 20/21"
  },
  "limits": [
    "Seven-task synthetic coding fixture with ambiguous pair representation; no general ranking",
    "No repository-level SWE result, long-context/vision qualification, or controlled performance population",
    "No fixed seed; cache warmth and co-resident background activity not fully controlled",
    "Only operator metadata redacted; complete native prompts, responses, reasoning, cases and grades are public",
    "Installed harness source revision unavailable; 1.5.0 collection differs from dirty 1.2.1 helper/publication worktree"
  ],
  "evidence_label": "functional",
  "default_recommendation": "keep-swift15",
  "configured_concurrency": 4,
  "request_concurrency": 1,
  "context_tokens": 32768,
  "restoration_verified": true
}