Unified Memory

Benchmarks / Which Small AI Model Should You Run on a Mac? We Ran 312 Tests / Summarize + extract

summarize-extract-017

hard extractionshipping

The prompt

user

Extract information from the email below. Reply with only a JSON object with exactly these keys:
- "order_number" (string): order number
- "tracking_number" (string): tracking number of the shipped package
- "arrival_date" (string "YYYY-MM-DD"): expected arrival date of the shipped package
- "shipped_items" (array): array of {"name": string, "quantity": number} for items that shipped
- "backordered_items" (array): array of {"name": string, "quantity": number} for items not yet shipped
- "order_total" (number): order total
- "charged_today" (number): amount charged today
- "backorder_tracking_number" (string): tracking number for the backordered package

Rules: include every key listed. Use null for any field that is not stated in the text (do not guess). Dates are YYYY-MM-DD. Times are HH:MM in 24-hour format. Amounts and counts are plain JSON numbers (no currency symbols, no commas). Phone numbers are in international format with + and the country code, for example +1 312 555 0100; leave out any national trunk prefix 0, so (0)20 or 020 becomes 20 after +44.

Email:
"""
From: Larkfield Home <hello@larkfieldhome.com>
Sent: Friday, December 4, 2026 11:52 AM
Subject: Part of your order is on its way (LH-30917)

Hi Dominique,

Good news - part of your order LH-30917 is on its way!

PACKAGE 1 of 2 - SHIPPED
Carrier: SwiftRoute
Tracking: SR 4410 9982 1137
Arriving: Tuesday
Contents: Linen Duvet Cover (Queen, Oat) x1, Linen Pillowcase Set (Oat) x2

PACKAGE 2 of 2 - BACKORDERED
Contents: Desk Lamp (Brass) x1
We expect this item to ship in about two weeks. We'll email tracking as soon as it leaves our warehouse.

Payment: Your order total is $214.85. We charged $189.90 today for the items that shipped. The remaining $24.95 will be charged when your lamp ships.
"""
What we expected
{
  "kind": "extract",
  "answer": {
    "order_number": "LH-30917",
    "tracking_number": "SR441099821137",
    "arrival_date": "2026-12-08",
    "shipped_items": [
      {
        "name": "Linen Duvet Cover (Queen, Oat)",
        "quantity": 1
      },
      {
        "name": "Linen Pillowcase Set (Oat)",
        "quantity": 2
      }
    ],
    "backordered_items": [
      {
        "name": "Desk Lamp (Brass)",
        "quantity": 1
      }
    ],
    "order_total": 214.85,
    "charged_today": 189.9,
    "backorder_tracking_number": null
  },
  "fields": {
    "order_number": {
      "type": "code"
    },
    "tracking_number": {
      "type": "code"
    },
    "arrival_date": {
      "type": "date"
    },
    "shipped_items": {
      "type": "list",
      "item_fields": {
        "name": {
          "type": "text",
          "needles_map": {
            "Linen Duvet Cover (Queen, Oat)": [
              "duvet"
            ],
            "Linen Pillowcase Set (Oat)": [
              "pillowcase",
              "pillow case"
            ]
          }
        },
        "quantity": {
          "type": "number"
        }
      }
    },
    "backordered_items": {
      "type": "list",
      "item_fields": {
        "name": {
          "type": "text",
          "needles_map": {
            "Desk Lamp (Brass)": [
              "lamp"
            ]
          }
        },
        "quantity": {
          "type": "number"
        }
      }
    },
    "order_total": {
      "type": "number"
    },
    "charged_today": {
      "type": "number"
    },
    "backorder_tracking_number": {
      "type": "code"
    }
  }
}

What each model answered

MiMo-V2.6 Distill 9B

✗ fail · score 88 · 5.28 s

{"order_number": "LH-30917", "tracking_number": "SR 4410 9982 1137", "arrival_date": null, "shipped_items": [{"name": "Linen Duvet Cover (Queen, Oat)", "quantity": 1}, {"name": "Linen Pillowcase Set (Oat)", "quantity": 2}], "backordered_items": [{"name": "Desk Lamp (Brass)", "quantity": 1}], "order_total": 214.85, "charged_today": 189.9, "backorder_tracking_number": null}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": null
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

Ornith-1.5 9B

✗ fail · score 88 · 5.09 s

{
  "order_number": "LH-30917",
  "tracking_number": "SR 4410 9982 1137",
  "arrival_date": null,
  "shipped_items": [
    {"name": "Linen Duvet Cover (Queen, Oat)", "quantity": 1},
    {"name": "Linen Pillowcase Set (Oat)", "quantity": 2}
  ],
  "backordered_items": [
    {"name": "Desk Lamp (Brass)", "quantity": 1}
  ],
  "order_total": 214.85,
  "charged_today": 189.9,
  "backorder_tracking_number": null
}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": null
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

Granite 4.2 8B

✗ fail · score 88 · 5.49 s

{
  "order_number": "LH-30917",
  "tracking_number": "SR 4410 9982 1137",
  "arrival_date": null,
  "shipped_items": [
    {
      "name": "Linen Duvet Cover (Queen, Oat)",
      "quantity": 1
    },
    {
      "name": "Linen Pillowcase Set (Oat)",
      "quantity": 2
    }
  ],
  "backordered_items": [
    {
      "name": "Desk Lamp (Brass)",
      "quantity": 1
    }
  ],
  "order_total": 214.85,
  "charged_today": 189.9,
  "backorder_tracking_number": null
}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": null
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

Spark-X2.5 4B

✗ fail · score 88 · 3.95 s

{
  "order_number": "LH-30917",
  "tracking_number": "SR 4410 9982 1137",
  "arrival_date": null,
  "shipped_items": [
    {
      "name": "Linen Duvet Cover (Queen, Oat)",
      "quantity": 1
    },
    {
      "name": "Linen Pillowcase Set (Oat)",
      "quantity": 2
    }
  ],
  "backordered_items": [
    {
      "name": "Desk Lamp (Brass)",
      "quantity": 1
    }
  ],
  "order_total": 214.85,
  "charged_today": 189.9,
  "backorder_tracking_number": null
}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": null
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

Mellum2.1 12B-A2.5B

✗ fail · score 88 · 2.69 s

{
  "order_number": "LH-30917",
  "tracking_number": "SR 4410 9982 1137",
  "arrival_date": "2026-12-13",
  "shipped_items": [
    {
      "name": "Linen Duvet Cover (Queen, Oat)",
      "quantity": 1
    },
    {
      "name": "Linen Pillowcase Set (Oat)",
      "quantity": 2
    }
  ],
  "backordered_items": [
    {
      "name": "Desk Lamp (Brass)",
      "quantity": 1
    }
  ],
  "order_total": 214.85,
  "charged_today": 189.9,
  "backorder_tracking_number": null
}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": "2026-12-13"
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

Ling-3.0-tiny (7.9B-A1.3B)

✓ pass · score 100 · 1.56 s

{"order_number": "LH-30917", "tracking_number": "SR 4410 9982 1137", "arrival_date": "2026-12-08", "shipped_items": [{"name": "Linen Duvet Cover", "quantity": 1}, {"name": "Linen Pillowcase Set", "quantity": 2}], "backordered_items": [{"name": "Desk Lamp", "quantity": 1}], "order_total": 214.85, "charged_today": 189.90, "backorder_tracking_number": null}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 1,
      "expected": "2026-12-08",
      "got": "2026-12-08"
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 8,
  "fields_total": 8
}

MiniCPM5 2B

✓ pass · score 100 · 1.76 s

{"order_number": "LH-30917", "tracking_number": "SR 4410 9982 1137", "arrival_date": "2026-12-08", "shipped_items": [{"name": "Linen Duvet Cover", "quantity": 1}, {"name": "Linen Pillowcase Set", "quantity": 2}], "backordered_items": [{"name": "Desk Lamp", "quantity": 1}], "order_total": 214.85, "charged_today": 189.90, "backorder_tracking_number": null}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 1,
      "expected": "2026-12-08",
      "got": "2026-12-08"
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 8,
  "fields_total": 8
}

K2-Horizon 0.9B

✗ fail · score 88 · 0.98 s

{
  "order_number": "LH-30917",
  "tracking_number": "SR 4410 9982 1137",
  "arrival_date": "2026-12-11",
  "shipped_items": [
    {
      "name": "Linen Duvet Cover (Queen, Oat)",
      "quantity": 1
    },
    {
      "name": "Linen Pillowcase Set (Oat)",
      "quantity": 2
    }
  ],
  "backordered_items": [
    {
      "name": "Desk Lamp (Brass)",
      "quantity": 1
    }
  ],
  "order_total": 214.85,
  "charged_today": 189.90,
  "backorder_tracking_number": null
}
Scoring detail
{
  "fields": {
    "order_number": {
      "score": 1,
      "expected": "LH-30917",
      "got": "LH-30917"
    },
    "tracking_number": {
      "score": 1,
      "expected": "SR441099821137",
      "got": "SR 4410 9982 1137"
    },
    "arrival_date": {
      "score": 0,
      "expected": "2026-12-08",
      "got": "2026-12-11"
    },
    "shipped_items": {
      "score": 1,
      "expected": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ],
      "got": [
        {
          "name": "Linen Duvet Cover (Queen, Oat)",
          "quantity": 1
        },
        {
          "name": "Linen Pillowcase Set (Oat)",
          "quantity": 2
        }
      ]
    },
    "backordered_items": {
      "score": 1,
      "expected": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ],
      "got": [
        {
          "name": "Desk Lamp (Brass)",
          "quantity": 1
        }
      ]
    },
    "order_total": {
      "score": 1,
      "expected": 214.85,
      "got": 214.85
    },
    "charged_today": {
      "score": 1,
      "expected": 189.9,
      "got": 189.9
    },
    "backorder_tracking_number": {
      "score": 1,
      "expected": null,
      "got": null
    }
  },
  "fields_correct": 7,
  "fields_total": 8
}

← summarize-extract-016All Summarize + extract itemssummarize-extract-018 →