{
  "schema_version": "iqai-public-ground-truth/1.0",
  "record_type": "Selected retained evaluation excerpt; not a live run or full export",
  "run_id": "LIVE-20261007T152856Z-11d243b8",
  "recorded_at_utc": "2026-10-07T15:28:56Z",
  "selection": "Previously retained presentation example: OpenAI baseline and all four follow-ups. All five baseline lanes are included for context. Not a random or representative sample.",
  "question": {
    "question_id": "q-number_theory_modular-0a1c60d0a065c57e",
    "operator_visible_question": "Compute the exact value of 97325^300 mod 916096. Give the exact non-negative residue.",
    "prompt": "Compute the exact value of 97325^300 mod 916096. Give the exact non-negative residue.\n\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>",
    "grading_rule": "EXACT_EQUIVALENCE",
    "tolerance": null
  },
  "reference": {
    "oracle_provider": "LOCAL_EXACT_MATH",
    "oracle_version": "local-exact-math/1.0.0",
    "methodology_version": "gt-oracle-methodology/1.0",
    "receipt_id": "oracle-local-174a7a7198dccde0",
    "receipt_hash": "174a7a7198dccde052f4dac62b9fa553daf08a80535a18bd7569b132ca642ca2",
    "retrieved_at_utc": "2026-10-07T15:28:56Z"
  },
  "reference_operation": "local_pow_mod",
  "reference_value": "746929",
  "protocol": {
    "depth": "STANDARD",
    "calls_per_provider": 5,
    "total_planned_calls": 25,
    "counterfactual_available": false
  },
  "selected_attempts": [
    {
      "stage": "Baseline",
      "attempt_id": "openai_gpt-5-2_BASELINE_0",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": "0",
      "reported_confidence": 99,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "89c39280ae81cb7f5a175bc681d61373d68a0eeeae13eeac2f52b4a161c3fdfc",
      "grading_receipt_id": "grade-f0cc5369c95ed542",
      "grading_receipt_hash": "f0cc5369c95ed5429bac9b7f5b106814c4c411f0943a401791ac5c80835eb0f3",
      "format_status": "COMPLIANT",
      "response_text": "Answer: 0\nConfidence: 99",
      "prompt_sent": "Compute the exact value of 97325^300 mod 916096. Give the exact non-negative residue.\n\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>"
    },
    {
      "stage": "Repeat 1",
      "attempt_id": "openai_gpt-5-2_REPEAT_1",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "REPEAT",
      "parent_attempt_id": null,
      "parsed_answer": "0",
      "reported_confidence": 100,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "7f422ddbde1dd681da8287d3091daf434d7b0c39c6e1b3c49e712fce04a85f00",
      "grading_receipt_id": "grade-d03f18fe3eadb972",
      "grading_receipt_hash": "d03f18fe3eadb9720da9a6268429975f9d2d0336841ac77c9e40ea7634944dbc",
      "format_status": "COMPLIANT",
      "response_text": "Answer: 0\nConfidence: 100",
      "prompt_sent": "Compute the exact value of 97325^300 mod 916096. Give the exact non-negative residue.\n\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>"
    },
    {
      "stage": "Repeat 2",
      "attempt_id": "openai_gpt-5-2_REPEAT_2",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "REPEAT",
      "parent_attempt_id": null,
      "parsed_answer": "0",
      "reported_confidence": 100,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "7f422ddbde1dd681da8287d3091daf434d7b0c39c6e1b3c49e712fce04a85f00",
      "grading_receipt_id": "grade-fdf58c2774741e56",
      "grading_receipt_hash": "fdf58c2774741e566c6014a74cf6688f002eb124c41bd6e7bc3683a4f5f45abe",
      "format_status": "COMPLIANT",
      "response_text": "Answer: 0\nConfidence: 100",
      "prompt_sent": "Compute the exact value of 97325^300 mod 916096. Give the exact non-negative residue.\n\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>"
    },
    {
      "stage": "Recheck",
      "attempt_id": "openai_gpt-5-2_INDEPENDENT_RECHECK_0",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "INDEPENDENT_RECHECK",
      "parent_attempt_id": "openai_gpt-5-2_BASELINE_0",
      "parsed_answer": "0",
      "reported_confidence": 99,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "89c39280ae81cb7f5a175bc681d61373d68a0eeeae13eeac2f52b4a161c3fdfc",
      "grading_receipt_id": "grade-1c0f9dff9016e48f",
      "grading_receipt_hash": "1c0f9dff9016e48f424ac3137c3a6e79cda2757bb2d5f36244eacc0d075fc7f6",
      "format_status": "COMPLIANT",
      "response_text": "Answer: 0\nConfidence: 99",
      "prompt_sent": "Recalculate the problem independently from the beginning. Do not assume your previous answer was correct. Return only:\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>"
    },
    {
      "stage": "Challenge",
      "attempt_id": "openai_gpt-5-2_NEUTRAL_CHALLENGE_0",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "NEUTRAL_CHALLENGE",
      "parent_attempt_id": "openai_gpt-5-2_BASELINE_0",
      "parsed_answer": "0",
      "reported_confidence": 99,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "89c39280ae81cb7f5a175bc681d61373d68a0eeeae13eeac2f52b4a161c3fdfc",
      "grading_receipt_id": "grade-c819767051f4dcab",
      "grading_receipt_hash": "c819767051f4dcab74ad22b0da506c8d4b34d452344d658833ae3ae16198eb2e",
      "format_status": "COMPLIANT",
      "response_text": "Answer: 0\nConfidence: 99",
      "prompt_sent": "Your previous answer may be incorrect. Check it carefully against the problem statement. Return only:\nRespond with exactly these two lines first (no derivation before them):\nAnswer: <final answer only>\nConfidence: <integer 0-100>"
    }
  ],
  "all_baselines": [
    {
      "attempt_id": "openai_gpt-5-2_BASELINE_0",
      "provider": "OPENAI",
      "model": "gpt-5.2",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": "0",
      "reported_confidence": 99,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "89c39280ae81cb7f5a175bc681d61373d68a0eeeae13eeac2f52b4a161c3fdfc",
      "grading_receipt_id": "grade-f0cc5369c95ed542",
      "grading_receipt_hash": "f0cc5369c95ed5429bac9b7f5b106814c4c411f0943a401791ac5c80835eb0f3",
      "format_status": "COMPLIANT"
    },
    {
      "attempt_id": "gemini_gemini-3-6-flash_BASELINE_0",
      "provider": "GEMINI",
      "model": "gemini-3.6-flash",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": "1",
      "reported_confidence": 100,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "4431fafb6d6fe7f93b8ca81d4fde0107ba9dccb2e850f00f0a49e3d8bfe76f23",
      "grading_receipt_id": "grade-7c485a88be4156bd",
      "grading_receipt_hash": "7c485a88be4156bdb6acf6fc39b26c8a28bedb27ccc5ba66dd632a7b870ec031",
      "format_status": "COMPLIANT"
    },
    {
      "attempt_id": "deepseek_deepseek-chat_BASELINE_0",
      "provider": "DEEPSEEK",
      "model": "deepseek-chat",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": "1",
      "reported_confidence": 100,
      "grading_status": "INCORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "4431fafb6d6fe7f93b8ca81d4fde0107ba9dccb2e850f00f0a49e3d8bfe76f23",
      "grading_receipt_id": "grade-1ce555ef3ee30f07",
      "grading_receipt_hash": "1ce555ef3ee30f072069b841a501c6ea458c92eba76b8639186606a26ef8e4c2",
      "format_status": "COMPLIANT"
    },
    {
      "attempt_id": "anthropic_claude-opus-4-6_BASELINE_0",
      "provider": "ANTHROPIC",
      "model": "claude-opus-4-6",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": null,
      "reported_confidence": null,
      "grading_status": "NOT_ASSESSED",
      "provider_status": "COMPLETED",
      "parsing_status": "FAILED",
      "response_hash": "768484150540e51d6c1826fce6ed9555aef649a4dba0eb2fe3c530540153308d",
      "grading_receipt_id": null,
      "grading_receipt_hash": null,
      "format_status": null
    },
    {
      "attempt_id": "xai_grok-3_BASELINE_0",
      "provider": "XAI",
      "model": "grok-3",
      "protocol_type": "BASELINE",
      "parent_attempt_id": null,
      "parsed_answer": "746929",
      "reported_confidence": 80,
      "grading_status": "CORRECT",
      "provider_status": "COMPLETED",
      "parsing_status": "PARSED",
      "response_hash": "001d6b99f8e1ee6b9e744d33a7921337651b4bb212c6c3e79db7531c77c67ea4",
      "grading_receipt_id": "grade-151de9f7010296bf",
      "grading_receipt_hash": "151de9f7010296bf9dcb2c7f8c041ae78006c3ae9fef4e7140220dcde63f6e95",
      "format_status": "COMPLIANT"
    }
  ],
  "boundaries": [
    "The reference receipt records local exact math, not Wolfram. Reference provenance must be checked per run.",
    "Reported confidence is the model's own statement, not a calibrated probability or evidence of correctness.",
    "Recheck and challenge are separate follow-ups to the original baseline using the same model; neither is the independent reference.",
    "Not assessed, provider failure, parsing failure, format compliance and wrong are distinct outcomes.",
    "This selected historical item is not a provider ranking, calibration study or general accuracy estimate.",
    "No model or provider calls were made to prepare this public excerpt."
  ]
}