Introduction: the 7B vision-language edition
Qwen2-VL-7B-Instruct is the instruction-tuned 7B member of Qwen2-VL. It combines text instructions with images or video and produces textual answers. Useful outputs include a page explanation, a comparison of pictures, a description of visible events or a response to a question about a chart. It remains this exact Qwen2-VL release; newer Qwen2.5-VL and Qwen3-VL checkpoints have their own capabilities and identities.
What Qwen2-VL adds
Qwen describes visual understanding across different image resolutions and aspect ratios, plus video question answering, dialogue and content creation. The family’s documented video capability includes material longer than 20 minutes. Duration alone does not establish how many frames an application processes or guarantee recall of every event. Multilingual image-text understanding includes English and Chinese, many European languages, Japanese, Korean, Arabic and Vietnamese.
The maker also describes integration with phones, robots and other devices through visual observations and text instructions. That is an application workflow: the model interprets the visual environment and helps choose actions, while connected software or hardware carries them out. Recognizing a screen does not independently give it control of the screen.
Architecture changes
Dynamic resolution converts an image into a variable number of visual tokens rather than forcing every input into the same representation. M-ROPE separates positional information for text, two-dimensional images and three-dimensional video information, combining spatial and temporal structure. The published 7B configuration lists 32,768 positions. Text, visual tokens and generated output share the finite sequence budget, so image resolution and video sampling affect the amount of material a request can include.
Preparing images and conversations
The official examples support a single image, several images, and text-only conversations. Attach the image with a specific question rather than relying on a broad instruction such as “analyze everything.” For a document page, ask for named fields and preserve punctuation in quotations. For two design screenshots, label them first and second, then request a table of visible changes. For a chart, ask for its labels, units and the values supporting a stated trend. These are practical applications of visual question answering, not promises of error-free extraction.
The maker’s utilities accept images from local files, encoded image data or image locations; its documented video example uses local video files or an ordered frame list. Application upload formats remain a separate interface decision. Preserve the original frame order and state the time interval represented by a frame list. The examples generate text; this checkpoint is not described as an image or video generator.
Resolution and video preparation
Qwen’s processor uses a default range of four to 16,384 visual tokens per image. It exposes min_pixels and max_pixels to balance detail against memory and computation, with a documented example targeting 256–1,280 visual tokens. Images are resized within the chosen bounds while keeping their aspect ratio. Explicit resized dimensions are rounded to a multiple of 28. A dense table or small writing may need enough resolution to remain readable; increasing it also consumes more of the available visual and context budget.
The video example shows a sampling rate of one frame per second and a configurable pixel bound. Those are example settings, not a universal required rate or hard media limit. For an event review, provide the relevant clip and request an ordered account of visible changes. Closely spaced events may require finer sampling than a broad scene summary. Qwen2-VL does not understand audio information inside videos, so spoken dialogue or sound effects require a separate audio workflow.
Known limitations
Qwen records an image-training-data cutoff of June 2023, with no assurance of knowledge about later people, products or events. Recognition of specific individuals and intellectual properties is incomplete. Complex multi-step instructions, object counts in crowded scenes and spatial relationships—especially in three dimensions—are documented weaknesses. Split a complicated visual task into clear questions, request the visible evidence supporting the result, and check extracted numbers and quoted text against the image. A plausible caption should not be treated as a precise count or a measured three-dimensional location.
The card’s short generation examples do not establish a separate hard output ceiling. Plan the answer size alongside text and visual input, and preserve enough detail in the source for the particular question. This makes model selection and input preparation more useful than assuming that any picture or any long video yields an equally reliable answer.
", "aboutText": "Introduction: the 7B vision-language edition\n\nQwen2-VL-7B-Instruct is the instruction-tuned 7B member of Qwen2-VL. It combines text instructions with images or video and produces textual answers. Useful outputs include a page explanation, a comparison of pictures, a description of visible events or a response to a question about a chart. It remains this exact Qwen2-VL release; newer Qwen2.5-VL and Qwen3-VL checkpoints have their own capabilities and identities.\n\nWhat Qwen2-VL adds\n\nQwen describes visual understanding across different image resolutions and aspect ratios, plus video question answering, dialogue and content creation. The family’s documented video capability includes material longer than 20 minutes. Duration alone does not establish how many frames an application processes or guarantee recall of every event. Multilingual image-text understanding includes English and Chinese, many European languages, Japanese, Korean, Arabic and Vietnamese.\n\nThe maker also describes integration with phones, robots and other devices through visual observations and text instructions. That is an application workflow: the model interprets the visual environment and helps choose actions, while connected software or hardware carries them out. Recognizing a screen does not independently give it control of the screen.\n\nArchitecture changes\n\nDynamic resolution converts an image into a variable number of visual tokens rather than forcing every input into the same representation. M-ROPE separates positional information for text, two-dimensional images and three-dimensional video information, combining spatial and temporal structure. The published 7B configuration lists 32,768 positions. Text, visual tokens and generated output share the finite sequence budget, so image resolution and video sampling affect the amount of material a request can include.\n\nPreparing images and conversations\n\nThe official examples support a single image, several images, and text-only conversations. Attach the image with a specific question rather than relying on a broad instruction such as “analyze everything.” For a document page, ask for named fields and preserve punctuation in quotations. For two design screenshots, label them first and second, then request a table of visible changes. For a chart, ask for its labels, units and the values supporting a stated trend. These are practical applications of visual question answering, not promises of error-free extraction.\n\nThe maker’s utilities accept images from local files, encoded image data or image locations; its documented video example uses local video files or an ordered frame list. Application upload formats remain a separate interface decision. Preserve the original frame order and state the time interval represented by a frame list. The examples generate text; this checkpoint is not described as an image or video generator.\n\nResolution and video preparation\n\nQwen’s processor uses a default range of four to 16,384 visual tokens per image. It exposes min_pixels and max_pixels to balance detail against memory and computation, with a documented example targeting 256–1,280 visual tokens. Images are resized within the chosen bounds while keeping their aspect ratio. Explicit resized dimensions are rounded to a multiple of 28. A dense table or small writing may need enough resolution to remain readable; increasing it also consumes more of the available visual and context budget.\n\nThe video example shows a sampling rate of one frame per second and a configurable pixel bound. Those are example settings, not a universal required rate or hard media limit. For an event review, provide the relevant clip and request an ordered account of visible changes. Closely spaced events may require finer sampling than a broad scene summary. Qwen2-VL does not understand audio information inside videos, so spoken dialogue or sound effects require a separate audio workflow.\n\nKnown limitations\n\nQwen records an image-training-data cutoff of June 2023, with no assurance of knowledge about later people, products or events. Recognition of specific individuals and intellectual properties is incomplete. Complex multi-step instructions, object counts in crowded scenes and spatial relationships—especially in three dimensions—are documented weaknesses. Split a complicated visual task into clear questions, request the visible evidence supporting the result, and check extracted numbers and quoted text against the image. A plausible caption should not be treated as a precise count or a measured three-dimensional location.\n\nThe card’s short generation examples do not establish a separate hard output ceiling. Plan the answer size alongside text and visual input, and preserve enough detail in the source for the particular question. This makes model selection and input preparation more useful than assuming that any picture or any long video yields an equally reliable answer.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Compare transcribed dashboard changes\n\nCompare these dashboard descriptions supplied as text. First: heading 'Orders', filter 'All dates', total 240. Second: heading 'Orders', filter 'This week', total 38. Return a table of changed and unchanged fields, then explain why the totals alone do not establish a fall in orders. Use only the descriptions and keep the filter labels exact.A field-comparison table and a short explanation of the different date scopes.The result does not treat totals from All dates and This week as comparable periods.Qwen2-VL-7B-Instruct-ultraQwen2-VL-7B-Instruct is Qwen’s instruction-tuned vision-language model for text, images and video, returning text. Use it to ask questions about document pages, compare pictures or describe visual events, with dynamic image resolution and documented limitations in counting, spatial reasoning and understanding audio in video.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Introduction: the 7B vision-language edition
Qwen2-VL-7B-Instruct is the instruction-tuned 7B member of Qwen2-VL. It combines text instructions with images or video and produces textual answers. Useful outputs include a page explanation, a comparison of pictures, a description of visible events or a response to a question about a chart. It remains this exact Qwen2-VL release; newer Qwen2.5-VL and Qwen3-VL checkpoints have their own capabilities and identities.
What Qwen2-VL adds
Qwen describes visual understanding across different image resolutions and aspect ratios, plus video question answering, dialogue and content creation. The family’s documented video capability includes material longer than 20 minutes. Duration alone does not establish how many frames an application processes or guarantee recall of every event. Multilingual image-text understanding includes English and Chinese, many European languages, Japanese, Korean, Arabic and Vietnamese.
The maker also describes integration with phones, robots and other devices through visual observations and text instructions. That is an application workflow: the model interprets the visual environment and helps choose actions, while connected software or hardware carries them out. Recognizing a screen does not independently give it control of the screen.
Architecture changes
Dynamic resolution converts an image into a variable number of visual tokens rather than forcing every input into the same representation. M-ROPE separates positional information for text, two-dimensional images and three-dimensional video information, combining spatial and temporal structure. The published 7B configuration lists 32,768 positions. Text, visual tokens and generated output share the finite sequence budget, so image resolution and video sampling affect the amount of material a request can include.
Preparing images and conversations
The official examples support a single image, several images, and text-only conversations. Attach the image with a specific question rather than relying on a broad instruction such as “analyze everything.” For a document page, ask for named fields and preserve punctuation in quotations. For two design screenshots, label them first and second, then request a table of visible changes. For a chart, ask for its labels, units and the values supporting a stated trend. These are practical applications of visual question answering, not promises of error-free extraction.
The maker’s utilities accept images from local files, encoded image data or image locations; its documented video example uses local video files or an ordered frame list. Application upload formats remain a separate interface decision. Preserve the original frame order and state the time interval represented by a frame list. The examples generate text; this checkpoint is not described as an image or video generator.
Resolution and video preparation
Qwen’s processor uses a default range of four to 16,384 visual tokens per image. It exposes min_pixels and max_pixels to balance detail against memory and computation, with a documented example targeting 256–1,280 visual tokens. Images are resized within the chosen bounds while keeping their aspect ratio. Explicit resized dimensions are rounded to a multiple of 28. A dense table or small writing may need enough resolution to remain readable; increasing it also consumes more of the available visual and context budget.
The video example shows a sampling rate of one frame per second and a configurable pixel bound. Those are example settings, not a universal required rate or hard media limit. For an event review, provide the relevant clip and request an ordered account of visible changes. Closely spaced events may require finer sampling than a broad scene summary. Qwen2-VL does not understand audio information inside videos, so spoken dialogue or sound effects require a separate audio workflow.
Known limitations
Qwen records an image-training-data cutoff of June 2023, with no assurance of knowledge about later people, products or events. Recognition of specific individuals and intellectual properties is incomplete. Complex multi-step instructions, object counts in crowded scenes and spatial relationships—especially in three dimensions—are documented weaknesses. Split a complicated visual task into clear questions, request the visible evidence supporting the result, and check extracted numbers and quoted text against the image. A plausible caption should not be treated as a precise count or a measured three-dimensional location.
The card’s short generation examples do not establish a separate hard output ceiling. Plan the answer size alongside text and visual input, and preserve enough detail in the source for the particular question. This makes model selection and input preparation more useful than assuming that any picture or any long video yields an equally reliable answer.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits Qwen2-VL-7B-Instruct-ultra | 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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.00058 / 1K tokens Output · Output tokens $0.00142 / 1K tokens Input · Cache read $0 / 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": "Qwen2-VL-7B-Instruct-ultra", "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='Qwen2-VL-7B-Instruct-ultra',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='Qwen2-VL-7B-Instruct-ultra',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 Qwen2-VL-7B-Instruct-ultra
Use the method paths shown above.
Qwen2-VL-7B-Instruct-ultraQwen2-VL-7B-Instruct is Qwen’s instruction-tuned vision-language model for text, images and video, returning text. Use it to ask questions about document pages, compare pictures or describe visual events, with dynamic image resolution and documented limitations in counting, spatial reasoning and understanding audio in video.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Introduction: the 7B vision-language edition
Qwen2-VL-7B-Instruct is the instruction-tuned 7B member of Qwen2-VL. It combines text instructions with images or video and produces textual answers. Useful outputs include a page explanation, a comparison of pictures, a description of visible events or a response to a question about a chart. It remains this exact Qwen2-VL release; newer Qwen2.5-VL and Qwen3-VL checkpoints have their own capabilities and identities.
What Qwen2-VL adds
Qwen describes visual understanding across different image resolutions and aspect ratios, plus video question answering, dialogue and content creation. The family’s documented video capability includes material longer than 20 minutes. Duration alone does not establish how many frames an application processes or guarantee recall of every event. Multilingual image-text understanding includes English and Chinese, many European languages, Japanese, Korean, Arabic and Vietnamese.
The maker also describes integration with phones, robots and other devices through visual observations and text instructions. That is an application workflow: the model interprets the visual environment and helps choose actions, while connected software or hardware carries them out. Recognizing a screen does not independently give it control of the screen.
Architecture changes
Dynamic resolution converts an image into a variable number of visual tokens rather than forcing every input into the same representation. M-ROPE separates positional information for text, two-dimensional images and three-dimensional video information, combining spatial and temporal structure. The published 7B configuration lists 32,768 positions. Text, visual tokens and generated output share the finite sequence budget, so image resolution and video sampling affect the amount of material a request can include.
Preparing images and conversations
The official examples support a single image, several images, and text-only conversations. Attach the image with a specific question rather than relying on a broad instruction such as “analyze everything.” For a document page, ask for named fields and preserve punctuation in quotations. For two design screenshots, label them first and second, then request a table of visible changes. For a chart, ask for its labels, units and the values supporting a stated trend. These are practical applications of visual question answering, not promises of error-free extraction.
The maker’s utilities accept images from local files, encoded image data or image locations; its documented video example uses local video files or an ordered frame list. Application upload formats remain a separate interface decision. Preserve the original frame order and state the time interval represented by a frame list. The examples generate text; this checkpoint is not described as an image or video generator.
Resolution and video preparation
Qwen’s processor uses a default range of four to 16,384 visual tokens per image. It exposes min_pixels and max_pixels to balance detail against memory and computation, with a documented example targeting 256–1,280 visual tokens. Images are resized within the chosen bounds while keeping their aspect ratio. Explicit resized dimensions are rounded to a multiple of 28. A dense table or small writing may need enough resolution to remain readable; increasing it also consumes more of the available visual and context budget.
The video example shows a sampling rate of one frame per second and a configurable pixel bound. Those are example settings, not a universal required rate or hard media limit. For an event review, provide the relevant clip and request an ordered account of visible changes. Closely spaced events may require finer sampling than a broad scene summary. Qwen2-VL does not understand audio information inside videos, so spoken dialogue or sound effects require a separate audio workflow.
Known limitations
Qwen records an image-training-data cutoff of June 2023, with no assurance of knowledge about later people, products or events. Recognition of specific individuals and intellectual properties is incomplete. Complex multi-step instructions, object counts in crowded scenes and spatial relationships—especially in three dimensions—are documented weaknesses. Split a complicated visual task into clear questions, request the visible evidence supporting the result, and check extracted numbers and quoted text against the image. A plausible caption should not be treated as a precise count or a measured three-dimensional location.
The card’s short generation examples do not establish a separate hard output ceiling. Plan the answer size alongside text and visual input, and preserve enough detail in the source for the particular question. This makes model selection and input preparation more useful than assuming that any picture or any long video yields an equally reliable answer.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits Qwen2-VL-7B-Instruct-ultra | 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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.00058 / 1K tokens Output · Output tokens $0.00142 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "Qwen2-VL-7B-Instruct-ultra",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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': 'Qwen2-VL-7B-Instruct-ultra', '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': 'Qwen2-VL-7B-Instruct-ultra', '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 Qwen2-VL-7B-Instruct-ultra
Use the method paths shown above.
Technical example in English
Compare transcribed dashboard changes\n\nCompare these dashboard descriptions supplied as text. First: heading 'Orders', filter 'All dates', total 240. Second: heading 'Orders', filter 'This week', total 38. Return a table of changed and unchanged fields, then explain why the totals alone do not establish a fall in orders. Use only the descriptions and keep the filter labels exact.A field-comparison table and a short explanation of the different date scopes.The result does not treat totals from All dates and This week as comparable periods.