{
  "version": 1,
  "date": "2026-08-08",
  "tokenizer": "o200k_base (OpenAI tiktoken, offline cache)",
  "page_image_tokens": {
    "standard_tier": 1496,
    "high_resolution_tier": 4758
  },
  "web": [
    {
      "file": "web-github-markitdown.html",
      "sha256": "d9b8966cf39b3ac6",
      "raw_html_tokens": 143818,
      "markdown_tokens": 6670,
      "saving_pct": 95.4
    },
    {
      "file": "web-mdn-http-overview.html",
      "sha256": "1b61f0c066a22942",
      "raw_html_tokens": 61212,
      "markdown_tokens": 16911,
      "saving_pct": 72.4
    },
    {
      "file": "web-python-tutorial.html",
      "sha256": "0837a50fdc3bfe20",
      "raw_html_tokens": 21924,
      "markdown_tokens": 6180,
      "saving_pct": 71.8
    },
    {
      "file": "web-wikipedia-markdown.html",
      "sha256": "02293a094d183c32",
      "raw_html_tokens": 97305,
      "markdown_tokens": 17771,
      "saving_pct": 81.7
    },
    {
      "file": "web-wikipedia-switzerland.html",
      "sha256": "f6cfc287c68201a4",
      "raw_html_tokens": 542470,
      "markdown_tokens": 134106,
      "saving_pct": 75.3
    }
  ],
  "pdf": [
    {
      "file": "pdf-arxiv-2506.16035.pdf",
      "sha256": "7b40d68ad136798f",
      "pages": 21,
      "markdown_tokens": 14816,
      "native_upload_tokens_standard": 46232,
      "native_upload_tokens_highres": 114734,
      "saving_pct_standard": 68.0,
      "saving_pct_highres": 87.1
    },
    {
      "file": "pdf-keydock-tdd.pdf",
      "sha256": "7f4c9ddca1f21d1d",
      "pages": 18,
      "markdown_tokens": 6135,
      "native_upload_tokens_standard": 33063,
      "native_upload_tokens_highres": 91779,
      "saving_pct_standard": 81.4,
      "saving_pct_highres": 93.3
    },
    {
      "file": "pdf-milldown-tdd.pdf",
      "sha256": "65695821af6e6395",
      "pages": 20,
      "markdown_tokens": 9406,
      "native_upload_tokens_standard": 39326,
      "native_upload_tokens_highres": 104566,
      "saving_pct_standard": 76.1,
      "saving_pct_highres": 91.0
    },
    {
      "file": "pdf-pricing-research.pdf",
      "sha256": "638015ac7368402f",
      "pages": 9,
      "markdown_tokens": 4405,
      "native_upload_tokens_standard": 17869,
      "native_upload_tokens_highres": 47227,
      "saving_pct_standard": 75.3,
      "saving_pct_highres": 90.7
    }
  ],
  "xlsx": [
    {
      "file": "xlsx-directory.xlsx",
      "sha256": "1ddba83c2b74b7c8",
      "markdown_table_tokens": 3945,
      "csv_tokens": 3466,
      "csv_saving_vs_markdown_pct": 12.1
    },
    {
      "file": "xlsx-kpis.xlsx",
      "sha256": "6eb37b855b3fd9ad",
      "markdown_table_tokens": 5108,
      "csv_tokens": 3607,
      "csv_saving_vs_markdown_pct": 29.4
    },
    {
      "file": "xlsx-transactions.xlsx",
      "sha256": "c0aa047904d6d79b",
      "markdown_table_tokens": 15298,
      "csv_tokens": 11901,
      "csv_saving_vs_markdown_pct": 22.2
    }
  ],
  "docx_illustration": [
    {
      "file": "docx-python-tutorial.docx",
      "sha256": "0559f020d5287bd8",
      "markdown_tokens": 4916,
      "raw_ooxml_tokens": 83941,
      "vs_ooxml_pct": 94.1,
      "plain_text_tokens": 4901,
      "vs_plain_text_pct": -0.3,
      "note": "raw OOXML is the baseline we refuse to claim; plain text is the one that actually applies"
    },
    {
      "file": "docx-sample-memo.docx",
      "sha256": "5ce3ff2a363d237d",
      "markdown_tokens": 149,
      "raw_ooxml_tokens": 1127,
      "vs_ooxml_pct": 86.8,
      "plain_text_tokens": 122,
      "vs_plain_text_pct": -22.1,
      "note": "raw OOXML is the baseline we refuse to claim; plain text is the one that actually applies"
    },
    {
      "file": "docx-wikipedia-markdown.docx",
      "sha256": "684d36f516550e86",
      "markdown_tokens": 6193,
      "raw_ooxml_tokens": 192259,
      "vs_ooxml_pct": 96.8,
      "plain_text_tokens": 6079,
      "vs_plain_text_pct": -1.9,
      "note": "raw OOXML is the baseline we refuse to claim; plain text is the one that actually applies"
    }
  ],
  "summary": {
    "web_saving_pct_range": [
      71.8,
      95.4
    ],
    "pdf_saving_pct_range_standard_tier": [
      68.0,
      81.4
    ],
    "pdf_saving_pct_range_high_resolution_tier": [
      87.1,
      93.3
    ],
    "xlsx_csv_saving_pct_range": [
      12.1,
      29.4
    ]
  }
}