{
 "schema": "cain.e115.agi_benchmark_status.v1",
 "as_of": "2026-11-19",
 "statement": "CAIN-42 is governed AGI infrastructure; it is not yet AGI. No recognized AGI benchmark has been passed or scored at a competitive level. This table separates what was run from what was not, and why.",
 "status_vocabulary": {"RAN": "a scored run exists in a published bundle", "IN_PROGRESS": "running now, no result yet", "PLANNED": "runnable on this host, not yet run", "BLOCKED": "cannot run here; exact blocker given", "NOT_TESTED": "never attempted"},
 "benchmarks": [
  {"name": "AMA-Bench (long-horizon agent memory, ICML 2026)", "source": "arXiv:2602.22769; HF AMA-bench/AMA-bench (MIT)", "status": "IN_PROGRESS",
   "cain_run": "E115 Track A: official token-F1 on the real-world open-ended set (2,496 QA; dev 636 / sealed 1,860), memory arms M0-M4",
   "limit": "the paper's headline accuracy uses an LLM judge (Qwen3-32B or GPT-5.2) that cannot run here: judge accuracy BLOCKED; F1 only"},
  {"name": "ARC-AGI-2 (abstraction and reasoning)", "source": "github.com/arcprize/ARC-AGI-2 (Apache-2.0), public evaluation set 120 tasks", "status": "PLANNED",
   "cain_run": "none yet; planned with a governed local model after the E113 sealed run frees the CPU",
   "limit": "local 0.5B-1.5B CPU models; expected near-zero; the public set is not the private leaderboard"},
  {"name": "ARC-AGI-1", "source": "arcprize", "status": "NOT_TESTED", "cain_run": "none", "limit": ""},
  {"name": "Humanity's Last Exam", "source": "Center for AI Safety / Scale AI", "status": "BLOCKED", "cain_run": "none",
   "limit": "needs a frontier model; the configured API key has no credit; local 0.5B-1.5B models are not meaningful here"},
  {"name": "GPQA Diamond", "source": "Rein et al. 2023", "status": "NOT_TESTED", "cain_run": "none", "limit": ""},
  {"name": "GAIA (general assistant)", "source": "Mialon et al. 2023", "status": "BLOCKED", "cain_run": "none",
   "limit": "needs web/tool access plus a frontier model; no API credit"},
  {"name": "SWE-bench Verified", "source": "Jimenez et al.; OpenAI-verified subset", "status": "BLOCKED", "cain_run": "none",
   "limit": "needs container-isolated repository execution and a frontier model; no API credit"},
  {"name": "OSWorld / WebArena (computer and web agents)", "source": "Xie et al. 2024; Zhou et al. 2023", "status": "NOT_TESTED", "cain_run": "none", "limit": ""},
  {"name": "LongMemEval / LoCoMo (conversational memory)", "source": "arXiv:2410.10813; ACL 2024", "status": "BLOCKED", "cain_run": "none",
   "limit": "official scoring uses an LLM judge that cannot run here"},
  {"name": "GSM8K (grade-school math)", "source": "Cobbe et al. 2021", "status": "RAN", "cain_run": "E72 small sample: 1/10 with a local model",
   "limit": "10 items; tiny local model; not a benchmark result"}
 ],
 "internal_constructed_suites_not_agi_benchmarks": ["E75 single-function repair 8/8", "E93 procedure synthesis 369/510", "E95 composition 79/90", "E107/E108 hidden-rule discovery", "E110 transfer 4/5 pathways", "E113 collective vs individual (running)", "E114 research lab"]
}
