Qwen2.5-VL-72B-Instruct is the instruction-tuned 72B edition of Qwen2.5-VL, designed to understand visual material alongside text. Qwen describes recognition of objects, products and landmarks as well as interpretation of the text, charts, icons, graphics and layout inside images. Inputs combine a written question with one or more images or a video; output is text, including structured descriptions and extracted information. Compared with Qwen2-VL, the producer emphasises improved perception of time and image scale and a revised dynamic-resolution vision encoder using window attention.
Visual grounding is a separate capability from describing a picture. The model can return bounding boxes or points, with labels and attributes in JSON. Qwen2.5-VL uses coordinates at the image’s actual size scale rather than a normalised coordinate scheme. Name the objects to locate and give the required fields, such as bbox_2d and label, together with the coordinate interpretation. If an image is resized before analysis, keep that size information with the returned coordinates so they can be interpreted against the same image. Qwen’s examples combine detection with counting and object descriptions.
The producer highlights text recognition across languages, orientations and different kinds of scene. Its examples extract lines of text, associate text with locations and pull specified fields from receipts or forms. For a document brief, supply a clear scan, define the fields and say how absent or unreadable information should be represented. An illustrative result can be a structured record containing a document identifier, date and line items, with uncertainties kept distinct from what is visibly present. The supplied image provides the evidence for that extraction.
Qwen’s announcement introduces QwenVL HTML for document parsing with layout information. It demonstrates magazines, research papers, web pages and mobile screenshots, combining headings, paragraphs, images and formulas with bounding-box information. This is useful when reading order and page structure matter as much as the words. Request the representation explicitly: plain text for reading, JSON for selected fields or the producer’s layout-oriented HTML format. A scanned report can be organised into headings and paragraphs before a further instruction asks for a summary of a particular section.
For video, Qwen documents understanding of material longer than one hour, event search and temporal localisation down to seconds. Dynamic frame-rate training and absolute time encoding help represent the pace and position of events. Official examples identify the segment where a described action occurs and return a series of events with start times, end times and descriptions. Name the event to find and specify seconds or a timestamp format. A practical brief can ask for the relevant segments in a long demonstration, followed by a concise description of what happens in each.
The model is also presented as a visual agent for computer and mobile use: it can reason from a screen and direct tools towards a task. Grounding and screen understanding can identify a relevant control or the next intended interaction. An application supplies the screen, available actions and goal, executes a requested action and returns the new state. For example, the brief can ask for the control needed to reach a specified view, with a point or box identifying it. The model’s textual action description and the application’s executed interaction are separate stages.
Qwen’s processor allows control of image pixel counts. Its documented default range is 4 to 16,384 visual tokens per image; 256 to 1,280 tokens is an example setting, not a universal image limit. Resizing can preserve aspect ratio, and explicit dimensions are rounded to a multiple of 28. Higher resolution uses more computation and may preserve small text or dense diagram details. Match preparation to the question, keeping important labels legible and the relevant page or region clear. Video examples use frame sequences or a video with frame-rate information, which matters when an answer must contain timestamps.
The card’s current configuration is set to 32,768 context tokens. It describes a YaRN extension but warns that this can substantially affect temporal and spatial localisation and does not recommend it for those tasks. It separately discusses enlarging the position setting for long video. Hour-scale comprehension is a documented capability rather than a universal upload-duration or output-length cap. Choose a result that retains the information needed downstream: fields for a record, coordinates for grounding, layout for a document or timed descriptions for a video.
", "aboutText": "Qwen2.5-VL-72B-Instruct is the instruction-tuned 72B edition of Qwen2.5-VL, designed to understand visual material alongside text. Qwen describes recognition of objects, products and landmarks as well as interpretation of the text, charts, icons, graphics and layout inside images. Inputs combine a written question with one or more images or a video; output is text, including structured descriptions and extracted information. Compared with Qwen2-VL, the producer emphasises improved perception of time and image scale and a revised dynamic-resolution vision encoder using window attention.\n\nVisual grounding is a separate capability from describing a picture. The model can return bounding boxes or points, with labels and attributes in JSON. Qwen2.5-VL uses coordinates at the image’s actual size scale rather than a normalised coordinate scheme. Name the objects to locate and give the required fields, such as bbox_2d and label, together with the coordinate interpretation. If an image is resized before analysis, keep that size information with the returned coordinates so they can be interpreted against the same image. Qwen’s examples combine detection with counting and object descriptions.\n\nThe producer highlights text recognition across languages, orientations and different kinds of scene. Its examples extract lines of text, associate text with locations and pull specified fields from receipts or forms. For a document brief, supply a clear scan, define the fields and say how absent or unreadable information should be represented. An illustrative result can be a structured record containing a document identifier, date and line items, with uncertainties kept distinct from what is visibly present. The supplied image provides the evidence for that extraction.\n\nQwen’s announcement introduces QwenVL HTML for document parsing with layout information. It demonstrates magazines, research papers, web pages and mobile screenshots, combining headings, paragraphs, images and formulas with bounding-box information. This is useful when reading order and page structure matter as much as the words. Request the representation explicitly: plain text for reading, JSON for selected fields or the producer’s layout-oriented HTML format. A scanned report can be organised into headings and paragraphs before a further instruction asks for a summary of a particular section.\n\nFor video, Qwen documents understanding of material longer than one hour, event search and temporal localisation down to seconds. Dynamic frame-rate training and absolute time encoding help represent the pace and position of events. Official examples identify the segment where a described action occurs and return a series of events with start times, end times and descriptions. Name the event to find and specify seconds or a timestamp format. A practical brief can ask for the relevant segments in a long demonstration, followed by a concise description of what happens in each.\n\nThe model is also presented as a visual agent for computer and mobile use: it can reason from a screen and direct tools towards a task. Grounding and screen understanding can identify a relevant control or the next intended interaction. An application supplies the screen, available actions and goal, executes a requested action and returns the new state. For example, the brief can ask for the control needed to reach a specified view, with a point or box identifying it. The model’s textual action description and the application’s executed interaction are separate stages.\n\nQwen’s processor allows control of image pixel counts. Its documented default range is 4 to 16,384 visual tokens per image; 256 to 1,280 tokens is an example setting, not a universal image limit. Resizing can preserve aspect ratio, and explicit dimensions are rounded to a multiple of 28. Higher resolution uses more computation and may preserve small text or dense diagram details. Match preparation to the question, keeping important labels legible and the relevant page or region clear. Video examples use frame sequences or a video with frame-rate information, which matters when an answer must contain timestamps.\n\nThe card’s current configuration is set to 32,768 context tokens. It describes a YaRN extension but warns that this can substantially affect temporal and spatial localisation and does not recommend it for those tasks. It separately discusses enlarging the position setting for long video. Hour-scale comprehension is a documented capability rather than a universal upload-duration or output-length cap. Choose a result that retains the information needed downstream: fields for a record, coordinates for grounding, layout for a document or timed descriptions for a video.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Extract fields from a receipt transcription\n\nExtract a record from this receipt transcription: “Harbor Stationery; receipt HS-81; 14 March; notebooks 3 x 4.00; total 12.00.” Return merchant, receipt_id, date_text, items, total and currency. Preserve the written date and mark currency as unknown because no symbol or currency name is present. Keep a supporting source phrase with each field.A field record with line-item details, preserved date text, source phrases and an explicit unknown currency.Check that the answer does not infer a year or currency and that item arithmetic agrees with the stated total.qwen/qwen2.5-vl-72b-instructQwen2.5-VL-72B-Instruct is Qwen’s vision-language model for images, documents, charts and video. It returns text, structured extracted data and visual coordinates, and can locate events in long video or reason about a computer or mobile screen.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen2.5-VL-72B-Instruct is the instruction-tuned 72B edition of Qwen2.5-VL, designed to understand visual material alongside text. Qwen describes recognition of objects, products and landmarks as well as interpretation of the text, charts, icons, graphics and layout inside images. Inputs combine a written question with one or more images or a video; output is text, including structured descriptions and extracted information. Compared with Qwen2-VL, the producer emphasises improved perception of time and image scale and a revised dynamic-resolution vision encoder using window attention.
Visual grounding is a separate capability from describing a picture. The model can return bounding boxes or points, with labels and attributes in JSON. Qwen2.5-VL uses coordinates at the image’s actual size scale rather than a normalised coordinate scheme. Name the objects to locate and give the required fields, such as bbox_2d and label, together with the coordinate interpretation. If an image is resized before analysis, keep that size information with the returned coordinates so they can be interpreted against the same image. Qwen’s examples combine detection with counting and object descriptions.
The producer highlights text recognition across languages, orientations and different kinds of scene. Its examples extract lines of text, associate text with locations and pull specified fields from receipts or forms. For a document brief, supply a clear scan, define the fields and say how absent or unreadable information should be represented. An illustrative result can be a structured record containing a document identifier, date and line items, with uncertainties kept distinct from what is visibly present. The supplied image provides the evidence for that extraction.
Qwen’s announcement introduces QwenVL HTML for document parsing with layout information. It demonstrates magazines, research papers, web pages and mobile screenshots, combining headings, paragraphs, images and formulas with bounding-box information. This is useful when reading order and page structure matter as much as the words. Request the representation explicitly: plain text for reading, JSON for selected fields or the producer’s layout-oriented HTML format. A scanned report can be organised into headings and paragraphs before a further instruction asks for a summary of a particular section.
For video, Qwen documents understanding of material longer than one hour, event search and temporal localisation down to seconds. Dynamic frame-rate training and absolute time encoding help represent the pace and position of events. Official examples identify the segment where a described action occurs and return a series of events with start times, end times and descriptions. Name the event to find and specify seconds or a timestamp format. A practical brief can ask for the relevant segments in a long demonstration, followed by a concise description of what happens in each.
The model is also presented as a visual agent for computer and mobile use: it can reason from a screen and direct tools towards a task. Grounding and screen understanding can identify a relevant control or the next intended interaction. An application supplies the screen, available actions and goal, executes a requested action and returns the new state. For example, the brief can ask for the control needed to reach a specified view, with a point or box identifying it. The model’s textual action description and the application’s executed interaction are separate stages.
Qwen’s processor allows control of image pixel counts. Its documented default range is 4 to 16,384 visual tokens per image; 256 to 1,280 tokens is an example setting, not a universal image limit. Resizing can preserve aspect ratio, and explicit dimensions are rounded to a multiple of 28. Higher resolution uses more computation and may preserve small text or dense diagram details. Match preparation to the question, keeping important labels legible and the relevant page or region clear. Video examples use frame sequences or a video with frame-rate information, which matters when an answer must contain timestamps.
The card’s current configuration is set to 32,768 context tokens. It describes a YaRN extension but warns that this can substantially affect temporal and spatial localisation and does not recommend it for those tasks. It separately discusses enlarging the position setting for long video. Hour-scale comprehension is a documented capability rather than a universal upload-duration or output-length cap. Choose a result that retains the information needed downstream: fields for a record, coordinates for grounding, layout for a document or timed descriptions for a video.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen2.5-vl-72b-instruct | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits Minimum items: 1 | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits Examples: system, user, assistant, tool, function Conditions: These roles are explicitly represented by Source compatibility types; the exact model role set remains model-specific. Model-specific | Default— |
Fieldmessages[].content | Typestring|content_part[]|null | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_calls | Typetool_call[] | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_call_id | Typestring | RequirementOptional | Values & limits Applies when: role=tool; use actual returned ID | Default— |
Fieldmessages[].name | Typestring | RequirementOptional | Values & limits | Default— |
Fieldmessages[].reasoning_content | Typestring | RequirementOptional | Values & limits Applies when: Round-trip actual returned reasoning where this model uses it | Default— |
Fieldmessages[].reasoning_details | Typeobject|array | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmax_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldmax_completion_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Fieldstream_options.include_usage | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtemperature | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtop_p | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldstop | Typestring|string[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools | Typefunction_tool[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools[].type | Typestring | RequirementOptional | Values & limits function Model-specific | Default— |
Fieldtools[].function.name | Typestring | RequirementRequired | Values & limits Applies when: function tool supplied | Default— |
Fieldtools[].function.description | Typestring | RequirementOptional | Values & limits | Default— |
Fieldtools[].function.parameters | TypeJSON_schema | RequirementOptional | Values & limits | Default— |
Fieldtool_choice | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldparallel_tool_calls | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldresponse_format | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning_effort | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldservice_tier | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldinstructions | Typestring | RequirementOptional | Values & limits Compatibility field | Default— |
Fieldfunctions | Typefunction[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldfunction_call | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldprompt_cache_options | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmodalities | Typestring[] | RequirementOptional | Values & limits text | Default— |
Fieldprovider | Typeobject | RequirementOptional | Values & limits Conditions: provider.options is rejected on this Ultra text route | Default— |
Fieldcache_control.ttl | Typestring | RequirementOptional | Values & limits Conditions: 1h is rejected in actual prompt-cache locations | Default— |
Text · UTF8 text
messages[].content
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages |
Parametermaterial_types | Contracttext |
Parametermessages | Contractmessage array with role/content shape |
Parameterfile_material | Contractactual UTF-8 document inserted as literal text; no file ID inferred |
Parameteroptional_controls | Contractstream, temperature, top_p, stop, max_tokens, max_completion_tokens, reasoning, response_format, tools, tool_choice |
Parametercontrol_scope | Contractmodel-specific; user function tools differ from hosted tools |
Parameteroutput_modalities | Contracttext |
Parameterhosted_operations | Contractfalse |
Parametermultimodal_input_binding | Contractnull |
Lifecycle: synchronous_or_streamed
Send exact model ID with messages.
Context length is not an output cap.
A text file may supply literal text; image/audio/file-object inputs need exact binding.
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0016 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0.0008 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
| ConfigurationRequest-length tier 1 | Rate schedule— | Context tokens, more than0 | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0016 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0.0008 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "qwen/qwen2.5-vl-72b-instruct", "messages": [{"role": "user", "content": "Explain the difference between an observation and a hypothesis in three sentences."}]}')\n\nPath("request.json").write_text(json.dumps(p,ensure_ascii=False))\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/chat/completions \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\nr=json.loads(Path("result.json").read_text())\nfor x in r.get("choices",[]):\n if x.get("message",{}).get("content"): print(x["message"]["content"])\nPY_RESULT\nfrom openai import OpenAI\nimport os\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nresult=client.chat.completions.create(model='qwen/qwen2.5-vl-72b-instruct',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}])\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nsave_result(result.to_dict(mode="json"))\nfrom openai import OpenAI\nimport os\nfrom pathlib import Path\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nwith client.chat.completions.stream(model='qwen/qwen2.5-vl-72b-instruct',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}]) as stream:\n with Path("stream-events.jsonl").open("w",encoding="utf-8") as log:\n for event in stream:\n payload=event.to_dict(mode="json")\n log.write(json.dumps(payload,ensure_ascii=False)+"\\n"); log.flush()\n print(json.dumps(payload,ensure_ascii=False))\n final=stream.get_final_completion()\n save_result(final.to_dict(mode="json"),"stream-result")\njson or SSE
Keep every complete returned response, file and content part.
Keep every choice and its content/tool calls/reasoning. In SSE assemble deltas by choice/tool index, retain the final finish reason and usage.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldchoices[].index | Typeinteger |
Fieldchoices[].message.content | Typestring|null |
Fieldchoices[].message.tool_calls | Typetool_call[] |
Fieldchoices[].message.reasoning_content | Typestring |
Fieldchoices[].message.reasoning_details | Typeobject|array |
Fieldchoices[].finish_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen2.5-vl-72b-instruct
Use the method paths shown above.
qwen/qwen2.5-vl-72b-instructQwen2.5-VL-72B-Instruct is Qwen’s vision-language model for images, documents, charts and video. It returns text, structured extracted data and visual coordinates, and can locate events in long video or reason about a computer or mobile screen.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen2.5-VL-72B-Instruct is the instruction-tuned 72B edition of Qwen2.5-VL, designed to understand visual material alongside text. Qwen describes recognition of objects, products and landmarks as well as interpretation of the text, charts, icons, graphics and layout inside images. Inputs combine a written question with one or more images or a video; output is text, including structured descriptions and extracted information. Compared with Qwen2-VL, the producer emphasises improved perception of time and image scale and a revised dynamic-resolution vision encoder using window attention.
Visual grounding is a separate capability from describing a picture. The model can return bounding boxes or points, with labels and attributes in JSON. Qwen2.5-VL uses coordinates at the image’s actual size scale rather than a normalised coordinate scheme. Name the objects to locate and give the required fields, such as bbox_2d and label, together with the coordinate interpretation. If an image is resized before analysis, keep that size information with the returned coordinates so they can be interpreted against the same image. Qwen’s examples combine detection with counting and object descriptions.
The producer highlights text recognition across languages, orientations and different kinds of scene. Its examples extract lines of text, associate text with locations and pull specified fields from receipts or forms. For a document brief, supply a clear scan, define the fields and say how absent or unreadable information should be represented. An illustrative result can be a structured record containing a document identifier, date and line items, with uncertainties kept distinct from what is visibly present. The supplied image provides the evidence for that extraction.
Qwen’s announcement introduces QwenVL HTML for document parsing with layout information. It demonstrates magazines, research papers, web pages and mobile screenshots, combining headings, paragraphs, images and formulas with bounding-box information. This is useful when reading order and page structure matter as much as the words. Request the representation explicitly: plain text for reading, JSON for selected fields or the producer’s layout-oriented HTML format. A scanned report can be organised into headings and paragraphs before a further instruction asks for a summary of a particular section.
For video, Qwen documents understanding of material longer than one hour, event search and temporal localisation down to seconds. Dynamic frame-rate training and absolute time encoding help represent the pace and position of events. Official examples identify the segment where a described action occurs and return a series of events with start times, end times and descriptions. Name the event to find and specify seconds or a timestamp format. A practical brief can ask for the relevant segments in a long demonstration, followed by a concise description of what happens in each.
The model is also presented as a visual agent for computer and mobile use: it can reason from a screen and direct tools towards a task. Grounding and screen understanding can identify a relevant control or the next intended interaction. An application supplies the screen, available actions and goal, executes a requested action and returns the new state. For example, the brief can ask for the control needed to reach a specified view, with a point or box identifying it. The model’s textual action description and the application’s executed interaction are separate stages.
Qwen’s processor allows control of image pixel counts. Its documented default range is 4 to 16,384 visual tokens per image; 256 to 1,280 tokens is an example setting, not a universal image limit. Resizing can preserve aspect ratio, and explicit dimensions are rounded to a multiple of 28. Higher resolution uses more computation and may preserve small text or dense diagram details. Match preparation to the question, keeping important labels legible and the relevant page or region clear. Video examples use frame sequences or a video with frame-rate information, which matters when an answer must contain timestamps.
The card’s current configuration is set to 32,768 context tokens. It describes a YaRN extension but warns that this can substantially affect temporal and spatial localisation and does not recommend it for those tasks. It separately discusses enlarging the position setting for long video. Hour-scale comprehension is a documented capability rather than a universal upload-duration or output-length cap. Choose a result that retains the information needed downstream: fields for a record, coordinates for grounding, layout for a document or timed descriptions for a video.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen2.5-vl-72b-instruct | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits user, assistant | Default— |
Fieldmessages[].content | Typestring | RequirementRequired | Values & limits | Default— |
Fieldmax_tokens | Typeinteger | RequirementRequired | Values & limits | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Text · UTF-8 text
messages[].content string
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages, max_tokens |
Parameteroptional_controls | Contractstream |
Parametermaterial_types | Contracttext |
Parametertext_file_material | ContractRead the actual UTF-8 document and put its text in the indicated input or message content string. |
Lifecycle: synchronous_or_streamed
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0016 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0.0008 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
| ConfigurationRequest-length tier 1 | Rate schedule— | Context tokens, more than0 | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0016 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0.0008 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen2.5-vl-72b-instruct",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\n: "${AILOOT_TEXT_FILE:?Set your actual UTF-8 document}"\nexport AILOOT_TEXT_FILE\npython3 - <<'PY_REQUEST'\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen2.5-vl-72b-instruct', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nPath('request.json').write_text(json.dumps(payload, ensure_ascii=False), encoding='utf-8')\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/messages \\\n -H "x-api-key: $AILOOT_API_KEY" -H 'anthropic-version: 2023-06-01' -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nPY_RESULT\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen2.5-vl-72b-instruct', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nfrom urllib.request import Request, urlopen\nfrom urllib.error import HTTPError\n\napi_key = os.environ['AILOOT_API_KEY']\nif not api_key.strip():\n raise SystemExit('Set AILOOT_API_KEY before making a request')\nheaders = {'x-api-key': 'PLACEHOLDER_KEY', 'anthropic-version': '2023-06-01', 'Content-Type': 'application/json'}\nheaders['x-api-key'] = api_key\nrequest = Request('https://api.ailoot.ai/v1/messages', data=json.dumps(payload, ensure_ascii=False).encode('utf-8'), headers=headers, method='POST')\ntry:\n with urlopen(request, timeout=120) as response:\n raw = response.read()\n Path('response-headers.txt').write_text(str(response.headers), encoding='utf-8')\nexcept HTTPError as error:\n Path('result.json').write_bytes(error.read())\n Path('response-headers.txt').write_text(str(error.headers), encoding='utf-8')\n raise\nPath('result.json').write_bytes(raw)\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nJSON or SSE
Keep every complete returned response, file and content part.
Keep all content blocks, thinking signatures and actual tool-use IDs. Assemble all SSE block deltas; a first text block is not the complete output.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldid | Typestring |
Fieldmodel | Typestring |
Fieldcontent[] | Typetext|thinking|tool_use|other content block |
Fieldstop_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen2.5-vl-72b-instruct
Use the method paths shown above.
Technical example in English
Extract fields from a receipt transcription\n\nExtract a record from this receipt transcription: “Harbor Stationery; receipt HS-81; 14 March; notebooks 3 x 4.00; total 12.00.” Return merchant, receipt_id, date_text, items, total and currency. Preserve the written date and mark currency as unknown because no symbol or currency name is present. Keep a supporting source phrase with each field.A field record with line-item details, preserved date text, source phrases and an explicit unknown currency.Check that the answer does not infer a year or currency and that item arithmetic agrees with the stated total.