{
  "benchmark": "paired_locomo_answer_codex_cli",
  "generated_at": "2026-09-25T15:40:48Z",
  "scope": {
    "source_questions": 1540,
    "evidence_bearing_questions": 1536,
    "evaluated_pairs": 10752,
    "cutoff": 50,
    "sampled": false
  },
  "protocol": {
    "answer_model": "gpt-6-astra",
    "judge_model": "gpt-6-astra",
    "answer_gold_visible": false,
    "answer_system_visible": false,
    "judge_memories_visible": false,
    "judge_system_visible": false,
    "same_questions_cutoff_prompts_and_models": true,
    "prior_compatible_blind_judgments_reused": 2086
  },
  "results": [
    { "rank": 1, "system": "Engram", "correct": 1377, "total": 1536, "accuracy": 89.65, "wilson_95": [88.02, 91.07] },
    { "rank": 2, "system": "Mem0 OSS 2.2.0", "correct": 1294, "total": 1536, "accuracy": 84.24, "wilson_95": [82.34, 85.98] },
    { "rank": 3, "system": "AgentMemory 0.9.29", "correct": 1280, "total": 1536, "accuracy": 83.33, "wilson_95": [81.39, 85.11] },
    { "rank": 4, "system": "LangMem / LangGraph 0.0.30", "correct": 1279, "total": 1536, "accuracy": 83.27, "wilson_95": [81.32, 85.05] },
    { "rank": 5, "system": "Session-file BM25", "correct": 1235, "total": 1536, "accuracy": 80.4, "wilson_95": [78.34, 82.31] },
    { "rank": 6, "system": "BM25", "correct": 1227, "total": 1536, "accuracy": 79.88, "wilson_95": [77.8, 81.81] },
    { "rank": 7, "system": "Tenure 1.0.36", "correct": 428, "total": 1536, "accuracy": 27.86, "wilson_95": [25.68, 30.16] }
  ],
  "pairwise_vs_engram": {
    "mem0": { "delta_points": 5.41, "engram_only_correct": 106, "competitor_only_correct": 23, "mcnemar_exact_p": 6.260285489771425e-14, "holm_adjusted_p": 6.260285489771425e-14 },
    "agentmemory": { "delta_points": 6.32, "engram_only_correct": 128, "competitor_only_correct": 31, "mcnemar_exact_p": 3.335166576122482e-15, "holm_adjusted_p": 6.670333152244964e-15 },
    "langmem": { "delta_points": 6.38, "engram_only_correct": 111, "competitor_only_correct": 13, "mcnemar_exact_p": 1.4575322761455106e-20, "holm_adjusted_p": 4.372596828436532e-20 },
    "session_bm25": { "delta_points": 9.25, "engram_only_correct": 184, "competitor_only_correct": 42, "mcnemar_exact_p": 2.175783334082806e-22, "holm_adjusted_p": 8.703133336331223e-22 },
    "bm25": { "delta_points": 9.77, "engram_only_correct": 175, "competitor_only_correct": 25, "mcnemar_exact_p": 6.55147784619449e-29, "holm_adjusted_p": 3.275738923097245e-28 },
    "tenure": { "delta_points": 61.79, "engram_only_correct": 960, "competitor_only_correct": 11, "mcnemar_exact_p": 1.7356642835102698e-267, "holm_adjusted_p": 1.041398570106162e-266 }
  },
  "retrieval_latency_note": "Engram used a hosted Voyage/Postgres path; competitor rows were local. Latency is disclosed in the report but is not a deployment-equivalent ranking.",
  "full_artifact_sha256": "a49a1fa8554199f9ab3c14252ec143f3a0fb3ef9677c39b903616d604e103fd9",
  "publication_boundary": "Engram-run diagnostic; not an official vendor submission or leaderboard result."
}