AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 20 November 2026

Does GPT-5.6 terra process a scanned page inside a mixed PDF?

GPT-5.6 terra converted the scanned page of a three-page report instead of skipping it. The photographed worksheet on page 2, which has no text layer, came back in its place between the two digital pages as tables, with values such as a flow cell pressure of 2.4 bar, 09:15 to 13:40 on site and the next service in January 2027 matching the page image.

1 of 1 test case passed

Every test case in this scenario has a result.

Pass rate100%1 of 1 gradable
Coverage1 of 1test case with a result
1PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the test case below: what we sent, what we checked, what the tool returned, and the proof.

The test case

Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.

Mixed pages — scanned page not skippedPassEvidence
What we sent

We sent the complete three-page PDF to the model as an input file in one API call, with the benchmark's conversion instruction, and asked for Markdown back.

InputDocument (PDF)
Input PDF: the three-page service report, with page 2 a photographed worksheet
Open the PDF ↗
What the tool returned
Markdown
# Equipment service report

## CLIENT

Braithe Estuary Partnership  
Estuary House, Vardenne Quay  
Braithe BR3 4LN

|  |  |
|---|---|
| Job reference | KAL-ESR-4471 |
| Date of visit | 29 January 2026 |
| Engineer | R. Tarbert |
| Instrument | Kestrel 400 turbidity analyser |
| Location | Vardenne Quay pumping station |
| Service interval | Twelve months |

## Reported fault

Intermittent baseline noise on channel 2, reported by the Partnership on 21 January. The analyser had been running without attention since the annual service in January 2025.

## Work carried out

The flow cell was removed, stripped and cleaned, and the pump tubing set replaced. The optical window was found to be lightly fouled on the inner face; it was cleaned rather than replaced, and should be inspected again at the next visit. The analyser was recalibrated against the two-point standards and left running on channel 2 for an hour before the site was handed back. The completed worksheet is bound in as page 2 of this report.

# Service worksheet

|  |  |
|---|---|
| Job reference | KAL-ESR-4471 |
| Instrument serial | GK-2214-77 |
| Engineer | R. Tarbert |
| Date | 29 January 2026 |

## CHECKS AND READINGS

|  |  |
|---|---|
| Flow cell condition on arrival | Fouled, inner face |
| Flow cell pressure after service | 2.4 bar |
| Zero drift over 30 minutes | 0.003 mg/l |
| Span recovery against standard | 99.4 per cent |
| Lamp hours at service | 8,140 |
| Result | Pass |

## Calibration certificate SR-4471/2

Analyser left in service on channel 2. Optical window to be inspected at the next visit.

## TIME ON SITE

|  |  |
|---|---|
| Arrived | 29 January 2026, 09:15 |
| Left | 29 January 2026, 13:40 |
| Chargeable time | 4 hours 25 minutes |
| Next service due | January 2027 |

## RECEIVED BY

|  |  |
|---|---|
| Name | M. Cairney |
| Position | Duty operator, Vardenne Quay |
| Date | 29 January 2026 |

[Image]

Engineer's signature

## Parts fitted

| Part | Code | Quantity |
|---|---|---|
| Pump tubing set | PT-114 | 1 |
| Flow cell O-ring, 12 mm | OR-0912 | 2 |
| Inlet filter, 40 micron | IF-4004 | 1 |

## Declaration

The work described in this report was carried out to the standard set out in the service agreement. The analyser was left in service and calibrated. The next annual service falls due in January 2027.

Engineer R. Tarbert

Countersigned Dr A. J. Prosser

Report issued 2 February 2026

Returned Markdown: the whole report, with the scanned worksheet between pages 1 and 3

Copied from Output file (Markdown), lines 1–87

Open the Markdown file ↗
Output file (JSON)
API response: the completed conversion, carrying the returned Markdown
Cited values1 value from this file
output 2 › content 1 › text
# Equipment service report ## CLIENT Braithe Estuary Partnership Estuary House, Vardenne Quay Braithe BR3 4LN | | | |---|---| | Job reference | KAL-ESR-4471 | | Date of visit | 29 January 2026 | | Engineer | R. Tarbert | | Instrument | Kestrel 400 turbidity analyser | | Location | Vardenne Quay pumping station | | Service interval | Twelve months | ## Reported fault Intermittent baseline noise on channel 2, reported by the Partnership on 21 January. The analyser had been running without attention since the annual service in January 2025. ## Work carried out The flow cell was removed, stripped and cleaned, and the pump tubing set replaced. The optical window was found to be lightly fouled on the inner face; it was cleaned rather than replaced, and should be inspected again at the next visit. The analyser was recalibrated against the two-point standards and left running on channel 2 for an hour before the site was handed back. The completed worksheet is bound in as page 2 of this report. # Service worksheet | | | |---|---| | Job reference | KAL-ESR-4471 | | Instrument serial | GK-2214-77 | | Engineer | R. Tarbert | | Date | 29 January 2026 | ## CHECKS AND READINGS | | | |---|---| | Flow cell condition on arrival | Fouled, inner face | | Flow cell pressure after service | 2.4 bar | | Zero drift over 30 minutes | 0.003 mg/l | | Span recovery against standard | 99.4 per cent | | Lamp hours at service | 8,140 | | Result | Pass | ## Calibration certificate SR-4471/2 Analyser left in service on channel 2. Optical window to be inspected at the next visit. ## TIME ON SITE | | | |---|---| | Arrived | 29 January 2026, 09:15 | | Left | 29 January 2026, 13:40 | | Chargeable time | 4 hours 25 minutes | | Next service due | January 2027 | ## RECEIVED BY | | | |---|---| | Name | M. Cairney | | Position | Duty operator, Vardenne Quay | | Date | 29 January 2026 | [Image] Engineer's signature ## Parts fitted | Part | Code | Quantity | |---|---|---| | Pump tubing set | PT-114 | 1 | | Flow cell O-ring, 12 mm | OR-0912 | 2 | | Inlet filter, 40 micron | IF-4004 | 1 | ## Declaration The work described in this report was carried out to the standard set out in the service agreement. The analyser was left in service and calibrated. The next annual service falls due in January 2027. Engineer R. Tarbert Countersigned Dr A. J. Prosser Report issued 2 February 2026
Show the whole file (8 KB)
JSON
{
  "id": "resp_00ccdb8c39b4d204006ab372deb0a887d19f4d924d626a57f7",
  "object": "response",
  "created_at": 1790145247,
  "status": "completed",
  "access_programs": null,
  "background": false,
  "billing": {
    "payer": "developer"
  },
  "completed_at": 1790145256,
  "error": null,
  "frequency_penalty": 0.0,
  "incomplete_details": null,
  "instructions": null,
  "max_output_tokens": null,
  "max_tool_calls": null,
  "model": "gpt-5.6-terra",
  "moderation": null,
  "output": [
    {
      "id": "rs_00ccdb8c39b4d204006ab372e03e2887d19a56ea63ce8678da",
      "type": "reasoning",
      "content": [],
      "encrypted_content": "gAAAAABqs3LoEupkgDrg-pNQInt3v0f200RwkiYiwLz4YdV7TfboVIutQB1OaaYm6NOx29aYF9mfMoHn37h4jm2EF7r5TtthGTipHEmn8bX7yepKAuBf5rqb9mPldxB0DfjmIfQ7hIwYln-J1TsX96CH5LNj3SJusbT1bbIuN--8QXQItFJibPskEGn5OAetNRt5WA1u50zkQ-G7xY76OGF_hqVljNF9g7eDsvDlbL5yEI-OdyPqVKG8CRP9_bSrtiLa0JUXRDWiwouuY_DeB7o-SoqfUHhJf-LY6yRHYBaLIzintUzbV7yf4CmxIL1JjDwFIlxX84DTTln4v8DVee6c9cBlYEDURDQzOu_FasC0GChEzd6HzNbqRwsDEbtbVtUAdonxXouyiwAejrPGihq3lxDqZPlMX-EYVkuZwYzoP-n4Y79lgExRyT7iX0AWhgcBCteN72YM_5rRnRE2GpRdpwo7KtFyjitGg0AKnnrsNxqBf6TbbHHRSZIWMHPVzNA5fO4PqbR8p1_AhcHTIH4PlQWZ_x0LdRBswCgJXe1H773AmQ7v8l4ympULSBmwgHqXBDw15mLR5sHDrcW2-vw3Z3gtCZOHh7ULGUVGrGhHYhDwZ0gTMUzzuy0w-Bjc7ruVp0ex0QvK7alkkoMVaVFBeFHmoB6BYrUJErG3Oew9bOnwWbVLIWaBx1AeVg33S3vapl11-0ggbWBRGT0sdL6HNjLkNhC847tKS_UW92wFuJMnnRSbyuVTH6c614ZE-eTD2rsYJCOdYbo3iFe_HDnEkiCRCiCfncNghAACk1c6u_AmKuZrqQKIKvOcsNhkA1aIjssPJzMo9cY-yamTjXLzkHtkYp3a02UWMVw66sl8zZHGxsQS23Nuc-8Sbbkn0s9X5BmxLggwq3lVMGigeRaZlrSoF2WHVH3IuutttpijXMlHXgd2LawQ09NemGyDZpA0J5PFHP6SgE5Bap8NPRJsf4ocwAIO2kqXgD78K1DT0u-BrrtWTj8Z1CyxBkYn2p2RdhRXJobR6u1JhZ1KVnWXTL4KJa9ZMG3pPrH-3NeNEyuXlKuWaLzQ8zLERp4mSwC0o_ksrjKMHq2b7ehPuH3HAn7Hn1V40hv8PaZM3Ds18G5wvIcU6OnXYZLZqsCpI2Os1C1zW0b7sYFm_N_m_4uBgT-Zq0x_IAGOgEfPhpm-Sfg1vfF1LebP-qlahaAUjfL4C7Zzl5oHKuAJLmduFWaHhDECI8O60-DN0QOoSFkwIcUHcQS_GJgjmPAD4HskncW_6gmqRcCyhk4t4WcfLyOqa4S-gUIZUrtZlNF3QmqsscqXpeGeKog36aXoh9z3kLc9KKBc9QNymjnYJ5KistQk7Vo0vY-sDXfLv8Opwt9koG-nGKEwMZbCvvZVExkPhxpMl5w84u6opuI52BsuD4OpI8_wBMKDPX9aXs4xRTQjuYsMGUKivbIPHyBedgyOU5QrQz6WidAzBDLmkrAa_pfRsKXgfCVlnqhG6K_UFesKTPcnSg0ZbtpjAG3L4YFMS8c6s7CK0ii5lMefW23a-IwpWKW2H8ZzvsmUVZEn8Ip1GpBel-EZQ0f2Z9QH4-_vc392pNkE1fnmG3i6AdphcJ1LNp2zg5f0_tBl3saFUuRT7yw-9u-nl2GxcswKFpyfewM7R_OEndafuBHI_gYRShyc8Dp1NTuF26pMZGVxGkxG94qjTdrKyKtseKPF3HQUKdYZIJSJhyKDKagl4xIs848BRfxgnsBEpLEw-lBhss9nbns3kejn8lDeouspv3TxvPKPV3Yiz5IeIO3tNaAsBsX44QqeqLhwcaUhuLlVgelI43gj-CO2T4w3omKVykP6OIhodg6IHQYmC_HGumjGNrrJsdupvgbZME53y7iG6r5SKOlCFjDKtbihdTdNzO7Gs3sc6zxLOs76d-4hqJr4Xwvx1qMCMKg1Eetok0BmkV0bNLnJj9vrmZ6czrHRPQJtWra95aAnToH0azFEUEAsbwovt5aRTQmoYqj4oeOtPYE4h5B5dy2lKOLRgUhWV8fQ1r4KV-F4UrKAPeUmcODNRL4_l6ZnYTP02I0WPk9k1CQsQ9Urby1YzOjCdL2XaJIjkFDABtvWeNJRQVYik0ZMP953KBpc9TqWxynYDGI1ACCqQrP9KKfdJPA42OCWCqf6Wvxogu2GXkxcPz4DLJaTffA0tt-5GBMYJqbr7wqH9JUBGh_WZqK5SnKb_kL9BYDy1W0_xFzPdOt7CuJMvXqwX1p6aPqf7PvsV18Z_mvyNZNsMIlElHp_DFXV2KafRbJpuJP15jxkrhNAyyc0LTJc42FCUiU4S_zLgjcsrv7JlErRtc8VlZHmrz2a6clo99xWXSPOjmzY0joChxOheGNuUd54bZTGG0B1dn01F5EJCtJeQWIMHwEwPRxtFoFe7ypLqRXSEXGmYaXDAvkVcx-U1VSliYgrDneaFet_15E_l_u2wGQZA93C6iwAMT4g3fiUG_tmX-hQxzRz3cYbXHvJXNbjNY0s9qtgPstbMN7lZPt8Uq50v2fgky7enkEdMIOD1cmUVXtjX1QV7P2I3N3Kf89toDo-bfWMea2YDymt5hbhrEVrCO-e7EZTOmMVEV51y5aEDEOPR3bYfz_Dyu1h49KKwLy1xtaEj_kOll9ckvnA43US7z9G226BhZAUWu9nnC0bAfasp59ycYBzsh2VIOeLuvq0aCbZM1d6E-sDhKwW3bF1TW_LxmvHLGdibWx1IYDbs626g3iDGVDNtGVnMLn0lClmvwAQ-jR9zVHUZRWvCabROXsjCJ1KBL2mqIgu6nKUsw_DlNHBBCZeprpEGwxGdwjkm_xK7B4AdGX40LMcH1MKD5TaLmIzRZUNl7ZyuxFL33krXoTLB3bL2LsGduiqsEjBrJyXOJq7ZPkIhItD998tXJqvOaNOFtYiKRnO42e1pjqMVlmKZ5qc5XR4XiKUdBtjHMLnDQOT_iLTNZdn3_5AdDZEFMK9KAgKmauNDKGZPco8ez0siZLPz59ih6On0UgBTDEgMGOQS2ZP3ORfANeacrk-zS45Mahet4ki9UBDUH3MrW3FZ0m7Qdjwj3_xQ6U5K86aWg==",
      "summary": []
    },
    {
      "id": "msg_00ccdb8c39b4d204006ab372e411a087d18878cc4c67435f42",
      "type": "message",
      "status": "completed",
      "content": [
        {
          "type": "output_text",
          "annotations": [],
          "logprobs": [],
          "text": "# Equipment service report\n\n## CLIENT\n\nBraithe Estuary Partnership  \nEstuary House, Vardenne Quay  \nBraithe BR3 4LN\n\n|  |  |\n|---|---|\n| Job reference | KAL-ESR-4471 |\n| Date of visit | 29 January 2026 |\n| Engineer | R. Tarbert |\n| Instrument | Kestrel 400 turbidity analyser |\n| Location | Vardenne Quay pumping station |\n| Service interval | Twelve months |\n\n## Reported fault\n\nIntermittent baseline noise on channel 2, reported by the Partnership on 21 January. The analyser had been running without attention since the annual service in January 2025.\n\n## Work carried out\n\nThe flow cell was removed, stripped and cleaned, and the pump tubing set replaced. The optical window was found to be lightly fouled on the inner face; it was cleaned rather than replaced, and should be inspected again at the next visit. The analyser was recalibrated against the two-point standards and left running on channel 2 for an hour before the site was handed back. The completed worksheet is bound in as page 2 of this report.\n\n# Service worksheet\n\n|  |  |\n|---|---|\n| Job reference | KAL-ESR-4471 |\n| Instrument serial | GK-2214-77 |\n| Engineer | R. Tarbert |\n| Date | 29 January 2026 |\n\n## CHECKS AND READINGS\n\n|  |  |\n|---|---|\n| Flow cell condition on arrival | Fouled, inner face |\n| Flow cell pressure after service | 2.4 bar |\n| Zero drift over 30 minutes | 0.003 mg/l |\n| Span recovery against standard | 99.4 per cent |\n| Lamp hours at service | 8,140 |\n| Result | Pass |\n\n## Calibration certificate SR-4471/2\n\nAnalyser left in service on channel 2. Optical window to be inspected at the next visit.\n\n## TIME ON SITE\n\n|  |  |\n|---|---|\n| Arrived | 29 January 2026, 09:15 |\n| Left | 29 January 2026, 13:40 |\n| Chargeable time | 4 hours 25 minutes |\n| Next service due | January 2027 |\n\n## RECEIVED BY\n\n|  |  |\n|---|---|\n| Name | M. Cairney |\n| Position | Duty operator, Vardenne Quay |\n| Date | 29 January 2026 |\n\n[Image]\n\nEngineer's signature\n\n## Parts fitted\n\n| Part | Code | Quantity |\n|---|---|---|\n| Pump tubing set | PT-114 | 1 |\n| Flow cell O-ring, 12 mm | OR-0912 | 2 |\n| Inlet filter, 40 micron | IF-4004 | 1 |\n\n## Declaration\n\nThe work described in this report was carried out to the standard set out in the service agreement. The analyser was left in service and calibrated. The next annual service falls due in January 2027.\n\nEngineer R. Tarbert\n\nCountersigned Dr A. J. Prosser\n\nReport issued 2 February 2026"
        }
      ],
      "phase": "final_answer",
      "role": "assistant"
    }
  ],
  "parallel_tool_calls": true,
  "presence_penalty": 0.0,
  "previous_response_id": null,
  "prompt_cache_key": null,
  "prompt_cache_retention": "24h",
  "reasoning": {
    "context": "all_turns",
    "effort": "medium",
    "mode": "standard",
    "summary": null
  },
  "safety_identifier": null,
  "service_tier": "default",
  "store": true,
  "temperature": 1.0,
  "text": {
    "format": {
      "type": "text"
    },
    "verbosity": "medium"
  },
  "tool_choice": "auto",
  "tool_usage": {
    "image_gen": {
      "input_tokens": 0,
      "input_tokens_details": {
        "image_tokens": 0,
        "text_tokens": 0
      },
      "output_tokens": 0,
      "output_tokens_details": {
        "image_tokens": 0,
        "text_tokens": 0
      },
      "total_tokens": 0
    },
    "web_search": {
      "num_requests": 0
    }
  },
  "tools": [],
  "top_logprobs": 0,
  "top_p": 0.98,
  "truncation": "disabled",
  "usage": {
    "input_tokens": 3681,
    "input_tokens_details": {
      "cache_write_tokens": 3678,
      "cached_tokens": 0
    },
    "output_tokens": 946,
    "output_tokens_details": {
      "reasoning_tokens": 273
    },
    "total_tokens": 4627
  },
  "user": null,
  "metadata": {}
}
Open the JSON file ↗
Expected vs. Found
✓Expected: The scanned page's text appears in the output the same way as the digital pages' text, and is not silently skipped.Found: The worksheet came back in place as tables, with all six checks and readings, including 2.4 bar and the result Pass; each value matches the page image.
Supporting proof
Proof 1Screenshot
Comparison: the scanned page 2 beside the returned Markdown, with the same tables and values
Open original ↗
Proof 2Screenshot
Run details: the setup, the complete PDF sent in one request, and the result
Open original ↗
Why this result

The scanned page had to come back like the digital pages, not be skipped. Page 2 is a photograph with no text layer, yet its worksheet appears in the returned Markdown in its place, every value matching the page image and the signature shown as a labelled placeholder.

Also observed in this test case

Worksheet letterhead left out. The worksheet's printed letterhead and its "FIELD SERVICE" label on the scanned page 2 do not appear in the returned Markdown.

Empty header rows on tables. Label-and-value blocks on both the digital pages and the scanned page come back as two-column tables with an empty header row, such as the job details at the top of the report.

Tested by Chandresh Bisht · evidence dated 23 September 2026

Configuration and setup

How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.

Software that produced the output
GPT-5.6 terra
Build or version
gpt-5.6-terra
Plan or tier
OpenAI API, credit-funded key (reference subject)
Surface
API
Set up before the run
A three-page service report: pages 1 and 3 digital with a text layer, page 2 a photographed worksheet with none. The complete PDF went in one request, under a fixed conversion instruction with default sampling, in one pass.
Conversion instruction
pdf2md-v1
Tested
23 September 2026 · Chandresh Bisht

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.

The rules
  • Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
  • Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
  • Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.

Where this sits in the benchmark

This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.

History of this result

What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.

from the publication record
20 November 2026First publishedConverting a complex PDF into clean Markdown with a hosted API v1

Act on this result

Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/pdf-to-markdown-apis/results/gpt-5-6-terra/a-document-mixing-digital-and-scanned-pages · 1 pass · 0 fail · coverage 1/1 · graded 2026-11-20

The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.

Verify the proof files

These files support this result. Open a file to inspect the original evidence.

Input · Document (PDF) ↗application/pdf · 356 KB
Output file (Markdown) ↗text/markdown; charset=utf-8 · 2 KB
Proof 1 · Screenshot ↗image/png · 1.2 MB
Output file (JSON) ↗application/json · 8 KB
Proof 2 · Screenshot ↗image/png · 68 KB
File fingerprints (SHA-256)

A fingerprint identifies the exact file used for this result.

Input · Document (PDF)096951b8eaf09740a926620943c7eb9b076e80144d76e7a90fb7f77056b6ef83
Output file (Markdown)fa4d09d00c62ce3530959ab9bcba20593639f7c11677b53fcc4c1ae9731706e9
Proof 1 · Screenshotadae6ff47428cff8694d54b0d475943faa94171d649bde4ae066829902207e3d
Output file (JSON)0b935418d6c1ad2edcd4971192c7515b6c33ab56196068ebeced1921497af330
Proof 2 · Screenshot9d0cbc991e845a1dcb14c1aa6ae7866e9d67f62771f95d90346a5386fc280b82