{
  "aggregates": {
    "case": {
      "eight-images-large": {
        "attempts": 1,
        "latency_p50_seconds": 28.170199,
        "latency_p95_seconds": 28.170199,
        "pass_rate": 1.0,
        "passed": 1
      }
    },
    "modality": {
      "image": {
        "attempts": 1,
        "latency_p50_seconds": 28.170199,
        "latency_p95_seconds": 28.170199,
        "pass_rate": 1.0,
        "passed": 1
      }
    }
  },
  "attempts": [
    {
      "case_id": "eight-images-large",
      "failure": null,
      "finish_reason": "stop",
      "media": [
        {
          "bytes": 44161,
          "kind": "image",
          "mime": "image/png",
          "path": "large-A.png",
          "sha256": "c9504138a9213677cb7e67189e015793edb5f55cbadf4fc476fb933aa309ee29"
        },
        {
          "bytes": 47789,
          "kind": "image",
          "mime": "image/png",
          "path": "large-B.png",
          "sha256": "9dfc9ddff04e3144d52303da0de15a70f05410982fbfd23801febbbfe71825d0"
        },
        {
          "bytes": 45978,
          "kind": "image",
          "mime": "image/png",
          "path": "large-C.png",
          "sha256": "8613b720c912aafc053b45d7d1bfd373c13fae64c374e60f3f0a86d477de6ef3"
        },
        {
          "bytes": 48918,
          "kind": "image",
          "mime": "image/png",
          "path": "large-D.png",
          "sha256": "562821d7f55a0fabc69da79f620ab57fd1c40c880ebc42b618f8c279e5981b36"
        },
        {
          "bytes": 48526,
          "kind": "image",
          "mime": "image/png",
          "path": "large-E.png",
          "sha256": "3c6b394442dc499ccbd2e73b4c27b7328bb3d7b07b081d4165d130a52f51f7bd"
        },
        {
          "bytes": 44484,
          "kind": "image",
          "mime": "image/png",
          "path": "large-F.png",
          "sha256": "f351432ea3e07dd34f0b3f7e8cc3f20f1a3c5576a5dfa1c3467fdf71a5eb9ac5"
        },
        {
          "bytes": 48820,
          "kind": "image",
          "mime": "image/png",
          "path": "large-G.png",
          "sha256": "3da5fd6fa55e33f07ce7228ddf3bc310a76dfa30e26f322dc100a2f7035c7874"
        },
        {
          "bytes": 47165,
          "kind": "image",
          "mime": "image/png",
          "path": "large-H.png",
          "sha256": "1be181e86a8e0867108112414968fb0c5154646af6d456cc48a25c4c8a01b488"
        }
      ],
      "modality": "image",
      "output": "CARD A: VALUE 17\nCARD B: VALUE 42\nCARD C: VALUE 9\nCARD D: VALUE 63\nCARD E: VALUE 28\nCARD F: VALUE 51\nCARD G: VALUE 34\nCARD H: VALUE 76\nMAX H 76\nMIN C 9\nSUM 320",
      "passed": true,
      "reasoning": {
        "chars": 342,
        "field": "reasoning",
        "tokens": null
      },
      "repetition": 1,
      "sampling": null,
      "seconds": 28.170199,
      "usage": {
        "completion_tokens": 242,
        "prompt_tokens": 63456,
        "prompt_tokens_details": null,
        "total_tokens": 63698
      },
      "validation": {
        "assertions": [
          {
            "passed": true,
            "type": "ordered_casefold",
            "values": [
              "CARD A",
              "17",
              "CARD B",
              "42",
              "CARD C",
              "9",
              "CARD D",
              "63",
              "CARD E",
              "28",
              "CARD F",
              "51",
              "CARD G",
              "34",
              "CARD H",
              "76"
            ]
          },
          {
            "passed": true,
            "type": "contains_casefold",
            "value": "320"
          },
          {
            "passed": true,
            "type": "contains_casefold",
            "value": "MAX H 76"
          },
          {
            "passed": true,
            "type": "contains_casefold",
            "value": "MIN C 9"
          }
        ],
        "finish_reason_allowed": true,
        "visible_answer": true
      }
    }
  ],
  "configuration": {
    "allowed_finish_reasons": [
      "stop"
    ],
    "chat_template_kwargs": null,
    "concurrency": 1,
    "max_images_per_request": 8,
    "max_tokens": 8192,
    "max_videos_per_request": 1,
    "reasoning_effort": "max",
    "sampling": {
      "temperature": {
        "effective_request": 0.0,
        "requested": null,
        "sent": true
      },
      "top_p": {
        "effective_request": null,
        "requested": null,
        "sent": false
      }
    },
    "thinking_mode": "default",
    "timeout_seconds": 180.0
  },
  "corpus": {
    "cases": [
      {
        "assertions": [
          {
            "type": "ordered_casefold",
            "values": [
              "CARD A",
              "17",
              "CARD B",
              "42",
              "CARD C",
              "9",
              "CARD D",
              "63",
              "CARD E",
              "28",
              "CARD F",
              "51",
              "CARD G",
              "34",
              "CARD H",
              "76"
            ]
          },
          {
            "type": "contains_casefold",
            "value": "320"
          },
          {
            "type": "contains_casefold",
            "value": "MAX H 76"
          },
          {
            "type": "contains_casefold",
            "value": "MIN C 9"
          }
        ],
        "id": "eight-images-large",
        "media": [
          {
            "bytes": 44161,
            "kind": "image",
            "mime": "image/png",
            "path": "large-A.png",
            "sha256": "c9504138a9213677cb7e67189e015793edb5f55cbadf4fc476fb933aa309ee29"
          },
          {
            "bytes": 47789,
            "kind": "image",
            "mime": "image/png",
            "path": "large-B.png",
            "sha256": "9dfc9ddff04e3144d52303da0de15a70f05410982fbfd23801febbbfe71825d0"
          },
          {
            "bytes": 45978,
            "kind": "image",
            "mime": "image/png",
            "path": "large-C.png",
            "sha256": "8613b720c912aafc053b45d7d1bfd373c13fae64c374e60f3f0a86d477de6ef3"
          },
          {
            "bytes": 48918,
            "kind": "image",
            "mime": "image/png",
            "path": "large-D.png",
            "sha256": "562821d7f55a0fabc69da79f620ab57fd1c40c880ebc42b618f8c279e5981b36"
          },
          {
            "bytes": 48526,
            "kind": "image",
            "mime": "image/png",
            "path": "large-E.png",
            "sha256": "3c6b394442dc499ccbd2e73b4c27b7328bb3d7b07b081d4165d130a52f51f7bd"
          },
          {
            "bytes": 44484,
            "kind": "image",
            "mime": "image/png",
            "path": "large-F.png",
            "sha256": "f351432ea3e07dd34f0b3f7e8cc3f20f1a3c5576a5dfa1c3467fdf71a5eb9ac5"
          },
          {
            "bytes": 48820,
            "kind": "image",
            "mime": "image/png",
            "path": "large-G.png",
            "sha256": "3da5fd6fa55e33f07ce7228ddf3bc310a76dfa30e26f322dc100a2f7035c7874"
          },
          {
            "bytes": 47165,
            "kind": "image",
            "mime": "image/png",
            "path": "large-H.png",
            "sha256": "1be181e86a8e0867108112414968fb0c5154646af6d456cc48a25c4c8a01b488"
          }
        ],
        "modality": "image",
        "prompt": "Read all eight images. Transcribe each in the supplied image order, exactly one line formatted CARD <letter>: VALUE <number>. Then add three lines: MAX <letter> <value>, MIN <letter> <value>, SUM <total>. Compute these across all eight images. No other text.",
        "repetitions": 1,
        "sampling": null
      }
    ],
    "path": "EVIDENCE_ROOT/corpus-large.json",
    "provenance": {
      "scope": "Eight distinct cards, ordered and reversed, independent exact OCR ordering and sum assertions.",
      "synthetic": {
        "license": "CC0-1.0"
      }
    },
    "schema": "multimodal-corpus/v1",
    "sha256": "8f808eef49c3d233c047904616768ce459a4015c977efe1bd6c47929eb3e8826"
  },
  "failed_attempts": 0,
  "finished_at": "2026-09-18T08:44:37Z",
  "identity": {
    "base_url": "http://127.0.0.1:39125/v1",
    "engine_build_ref": "6dc2f516688fe6f84c6994dcd20fddf296853a6c",
    "engine_revision": null,
    "hardware": "2x RTX PRO 6000 Max-Q 96GB TP2",
    "model": "glm53-flash-exl3-4bpw-v84-fp8-tp2-c4-327k-r9-vision8-nospec",
    "model_revision": "a5fee929cf4888b1824323e33e8a19b60129e025",
    "runtime_image": "verdictai/glm53-flash-exl3-k4@sha256:0f1cdcc8891f1cc3a444121eb61d366289a1cbba285f0892dcbb24bc94961692",
    "served_models": [
      "glm53-flash-exl3-4bpw-v84-fp8-tp2-c4-327k-r9-vision8-nospec"
    ]
  },
  "passed": true,
  "passed_attempts": 1,
  "requests": 1,
  "run_id": "multimodal-20260918T084408Z",
  "schema": "multimodal-benchmark-evidence/v1",
  "started_at": "2026-09-18T08:44:08Z"
}