The 30B-A3B Thinking edition
Qwen3-VL-30B-A3B-Thinking is the reasoning-enhanced mixture-of-experts edition of Qwen3-VL. It accepts text, images and video and produces text, including code or an explained conclusion. Keep its Thinking identity distinct from Qwen3-VL-30B-A3B-Instruct and other family sizes. The maker’s edition distinction describes the intended reasoning behavior without establishing an application-specific hybrid toggle.
Visual agents and coding
Qwen documents a visual-agent workflow that recognizes computer or mobile controls, understands their functions and invokes tools. For an interface task, provide the current screenshot and the intended goal, then request the relevant control and an explanation tied to the visible evidence. The connected application carries out actions and returns results; the model’s visual reasoning does not independently control the device.
The family also generates diagram content, HTML, CSS and JavaScript from images or video. Supply the visual reference with functional requirements and request editable source. For an existing interface, identify the intended change and relevant code rather than expect an unexplained picture to define every interaction. Generated source is an artifact to render and review, not proof of a working application.
Space, long context and video
The maker describes judgments about object positions, viewpoints and occlusions, with two-dimensional grounding and support for three-dimensional grounding. Ask which relationship matters and which cues support the conclusion. A visual relationship should remain distinct from a calibrated physical measurement.
The native context is 262,144 tokens. The family describes books and hours-long video, with second-level event indexing and expansion to about one million tokens through configuration. Preserve timing, frame order and source identifiers when asking for an event account. Context and sampling determine the material represented in a request; the family description should not become an unconditional promise of perfect recall.
Multimodal reasoning, recognition and OCR
The Thinking edition is relevant to STEM, mathematical and logical questions that connect visual evidence with an explanation. Provide the full statement, diagram labels and units and ask for the values and relationships used before the final answer. A clear deliverable can be a delimited result, evidence table or short conclusion backed by the image and text.
The family expands visual recognition across people, products, landmarks, plants, animals and media characters. OCR covers 32 languages and targets reading under low light, blur or tilt, including specialist vocabulary and document structure. Preserve legibility and request fields or quotations that can be checked. These use cases apply documented capabilities; they do not guarantee a correct identity, transcription or mathematical conclusion for every input.
Architecture
The exact configuration identifies a vision-language mixture-of-experts model with 48 text layers, 128 experts and eight experts selected per token. Interleaved-MRoPE distributes positional information across time, width and height. DeepStack combines visual features from several encoder levels, and text–timestamp alignment helps ground events in temporal references. Use this 30B-A3B Thinking edition consistently when preparing visual reasoning tasks and comparing their results.
Preparing inputs
The family’s examples support text-only tasks, single or multiple images and video. Label visual inputs and preserve their relationship. Separate image and video pixel budgets allow a workflow to control detail; explicit image dimensions are aligned to multiples of 32. The documented video default is two frames per second, with sampling-rate or frame-count controls where supported. Keep video metadata and avoid resizing twice when the helper has already resized the media.
For a document comparison, organize the result by page identifier. For a clip, request the relevant event, interval and evidence. Reserve sequence space for thinking and the final text as well as the visual tokens.
Generation guidance and limits
The checkpoint’s generation file defaults to temperature 0.8, top_p 0.95, top_k 20 and repetition penalty 1.0. The family’s separate Thinking evaluation recipe uses temperature 0.6, presence penalty zero, repetition penalty 1.0 and a 40,960-token output allowance. These are distinct producer presets; the allowance is not an independently established hard output maximum.
The family’s extension procedure uses YaRN and suggests a factor of two or three for the interleaved positional scheme when moving from 256K to about 1M. It should not inherit another family’s factor of four. Check extracted labels, calculations and generated code against the original evidence, and keep the final artifact distinguishable from the reasoning. The practical result is a reviewable visual conclusion or code proposal with this exact Thinking edition preserved.
", "aboutText": "The 30B-A3B Thinking edition\n\nQwen3-VL-30B-A3B-Thinking is the reasoning-enhanced mixture-of-experts edition of Qwen3-VL. It accepts text, images and video and produces text, including code or an explained conclusion. Keep its Thinking identity distinct from Qwen3-VL-30B-A3B-Instruct and other family sizes. The maker’s edition distinction describes the intended reasoning behavior without establishing an application-specific hybrid toggle.\n\nVisual agents and coding\n\nQwen documents a visual-agent workflow that recognizes computer or mobile controls, understands their functions and invokes tools. For an interface task, provide the current screenshot and the intended goal, then request the relevant control and an explanation tied to the visible evidence. The connected application carries out actions and returns results; the model’s visual reasoning does not independently control the device.\n\nThe family also generates diagram content, HTML, CSS and JavaScript from images or video. Supply the visual reference with functional requirements and request editable source. For an existing interface, identify the intended change and relevant code rather than expect an unexplained picture to define every interaction. Generated source is an artifact to render and review, not proof of a working application.\n\nSpace, long context and video\n\nThe maker describes judgments about object positions, viewpoints and occlusions, with two-dimensional grounding and support for three-dimensional grounding. Ask which relationship matters and which cues support the conclusion. A visual relationship should remain distinct from a calibrated physical measurement.\n\nThe native context is 262,144 tokens. The family describes books and hours-long video, with second-level event indexing and expansion to about one million tokens through configuration. Preserve timing, frame order and source identifiers when asking for an event account. Context and sampling determine the material represented in a request; the family description should not become an unconditional promise of perfect recall.\n\nMultimodal reasoning, recognition and OCR\n\nThe Thinking edition is relevant to STEM, mathematical and logical questions that connect visual evidence with an explanation. Provide the full statement, diagram labels and units and ask for the values and relationships used before the final answer. A clear deliverable can be a delimited result, evidence table or short conclusion backed by the image and text.\n\nThe family expands visual recognition across people, products, landmarks, plants, animals and media characters. OCR covers 32 languages and targets reading under low light, blur or tilt, including specialist vocabulary and document structure. Preserve legibility and request fields or quotations that can be checked. These use cases apply documented capabilities; they do not guarantee a correct identity, transcription or mathematical conclusion for every input.\n\nArchitecture\n\nThe exact configuration identifies a vision-language mixture-of-experts model with 48 text layers, 128 experts and eight experts selected per token. Interleaved-MRoPE distributes positional information across time, width and height. DeepStack combines visual features from several encoder levels, and text–timestamp alignment helps ground events in temporal references. Use this 30B-A3B Thinking edition consistently when preparing visual reasoning tasks and comparing their results.\n\nPreparing inputs\n\nThe family’s examples support text-only tasks, single or multiple images and video. Label visual inputs and preserve their relationship. Separate image and video pixel budgets allow a workflow to control detail; explicit image dimensions are aligned to multiples of 32. The documented video default is two frames per second, with sampling-rate or frame-count controls where supported. Keep video metadata and avoid resizing twice when the helper has already resized the media.\n\nFor a document comparison, organize the result by page identifier. For a clip, request the relevant event, interval and evidence. Reserve sequence space for thinking and the final text as well as the visual tokens.\n\nGeneration guidance and limits\n\nThe checkpoint’s generation file defaults to temperature 0.8, top_p 0.95, top_k 20 and repetition penalty 1.0. The family’s separate Thinking evaluation recipe uses temperature 0.6, presence penalty zero, repetition penalty 1.0 and a 40,960-token output allowance. These are distinct producer presets; the allowance is not an independently established hard output maximum.\n\nThe family’s extension procedure uses YaRN and suggests a factor of two or three for the interleaved positional scheme when moving from 256K to about 1M. It should not inherit another family’s factor of four. Check extracted labels, calculations and generated code against the original evidence, and keep the final artifact distinguishable from the reasoning. The practical result is a reviewable visual conclusion or code proposal with this exact Thinking edition preserved.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Reason about a transcribed geometry diagram\n\nUse this diagram transcription: rectangle ABCD has AB = 10 cm and BC = 6 cm; E is the midpoint of AB; segment CE is drawn. Find the area of triangle BCE. State the base and perpendicular height used, preserve the units, and give a short derivation. Do not assume any additional markings or image input.A short geometry derivation naming base BE and height BC, with the final area in square centimetres.The midpoint halves AB, and the triangle area uses one-half times the perpendicular base-height product.qwen/qwen3-vl-30b-a3b-thinkingQwen3-VL-30B-A3B-Thinking is Qwen’s mixture-of-experts vision-language edition for deeper text, image and video reasoning. It returns text or code for STEM questions, documents, spatial analysis and visual workflows, with 32-language OCR and a native 262,144-token context.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
The 30B-A3B Thinking edition
Qwen3-VL-30B-A3B-Thinking is the reasoning-enhanced mixture-of-experts edition of Qwen3-VL. It accepts text, images and video and produces text, including code or an explained conclusion. Keep its Thinking identity distinct from Qwen3-VL-30B-A3B-Instruct and other family sizes. The maker’s edition distinction describes the intended reasoning behavior without establishing an application-specific hybrid toggle.
Visual agents and coding
Qwen documents a visual-agent workflow that recognizes computer or mobile controls, understands their functions and invokes tools. For an interface task, provide the current screenshot and the intended goal, then request the relevant control and an explanation tied to the visible evidence. The connected application carries out actions and returns results; the model’s visual reasoning does not independently control the device.
The family also generates diagram content, HTML, CSS and JavaScript from images or video. Supply the visual reference with functional requirements and request editable source. For an existing interface, identify the intended change and relevant code rather than expect an unexplained picture to define every interaction. Generated source is an artifact to render and review, not proof of a working application.
Space, long context and video
The maker describes judgments about object positions, viewpoints and occlusions, with two-dimensional grounding and support for three-dimensional grounding. Ask which relationship matters and which cues support the conclusion. A visual relationship should remain distinct from a calibrated physical measurement.
The native context is 262,144 tokens. The family describes books and hours-long video, with second-level event indexing and expansion to about one million tokens through configuration. Preserve timing, frame order and source identifiers when asking for an event account. Context and sampling determine the material represented in a request; the family description should not become an unconditional promise of perfect recall.
Multimodal reasoning, recognition and OCR
The Thinking edition is relevant to STEM, mathematical and logical questions that connect visual evidence with an explanation. Provide the full statement, diagram labels and units and ask for the values and relationships used before the final answer. A clear deliverable can be a delimited result, evidence table or short conclusion backed by the image and text.
The family expands visual recognition across people, products, landmarks, plants, animals and media characters. OCR covers 32 languages and targets reading under low light, blur or tilt, including specialist vocabulary and document structure. Preserve legibility and request fields or quotations that can be checked. These use cases apply documented capabilities; they do not guarantee a correct identity, transcription or mathematical conclusion for every input.
Architecture
The exact configuration identifies a vision-language mixture-of-experts model with 48 text layers, 128 experts and eight experts selected per token. Interleaved-MRoPE distributes positional information across time, width and height. DeepStack combines visual features from several encoder levels, and text–timestamp alignment helps ground events in temporal references. Use this 30B-A3B Thinking edition consistently when preparing visual reasoning tasks and comparing their results.
Preparing inputs
The family’s examples support text-only tasks, single or multiple images and video. Label visual inputs and preserve their relationship. Separate image and video pixel budgets allow a workflow to control detail; explicit image dimensions are aligned to multiples of 32. The documented video default is two frames per second, with sampling-rate or frame-count controls where supported. Keep video metadata and avoid resizing twice when the helper has already resized the media.
For a document comparison, organize the result by page identifier. For a clip, request the relevant event, interval and evidence. Reserve sequence space for thinking and the final text as well as the visual tokens.
Generation guidance and limits
The checkpoint’s generation file defaults to temperature 0.8, top_p 0.95, top_k 20 and repetition penalty 1.0. The family’s separate Thinking evaluation recipe uses temperature 0.6, presence penalty zero, repetition penalty 1.0 and a 40,960-token output allowance. These are distinct producer presets; the allowance is not an independently established hard output maximum.
The family’s extension procedure uses YaRN and suggests a factor of two or three for the interleaved positional scheme when moving from 256K to about 1M. It should not inherit another family’s factor of four. Check extracted labels, calculations and generated code against the original evidence, and keep the final artifact distinguishable from the reasoning. The practical result is a reviewable visual conclusion or code proposal with this exact Thinking edition preserved.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-vl-30b-a3b-thinking | 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 $0.00058 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0 / 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 $0.00058 / 1K tokens Output · Output tokens $0.002 / 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": "qwen/qwen3-vl-30b-a3b-thinking", "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/qwen3-vl-30b-a3b-thinking',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/qwen3-vl-30b-a3b-thinking',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/qwen3-vl-30b-a3b-thinking
Use the method paths shown above.
qwen/qwen3-vl-30b-a3b-thinkingQwen3-VL-30B-A3B-Thinking is Qwen’s mixture-of-experts vision-language edition for deeper text, image and video reasoning. It returns text or code for STEM questions, documents, spatial analysis and visual workflows, with 32-language OCR and a native 262,144-token context.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
The 30B-A3B Thinking edition
Qwen3-VL-30B-A3B-Thinking is the reasoning-enhanced mixture-of-experts edition of Qwen3-VL. It accepts text, images and video and produces text, including code or an explained conclusion. Keep its Thinking identity distinct from Qwen3-VL-30B-A3B-Instruct and other family sizes. The maker’s edition distinction describes the intended reasoning behavior without establishing an application-specific hybrid toggle.
Visual agents and coding
Qwen documents a visual-agent workflow that recognizes computer or mobile controls, understands their functions and invokes tools. For an interface task, provide the current screenshot and the intended goal, then request the relevant control and an explanation tied to the visible evidence. The connected application carries out actions and returns results; the model’s visual reasoning does not independently control the device.
The family also generates diagram content, HTML, CSS and JavaScript from images or video. Supply the visual reference with functional requirements and request editable source. For an existing interface, identify the intended change and relevant code rather than expect an unexplained picture to define every interaction. Generated source is an artifact to render and review, not proof of a working application.
Space, long context and video
The maker describes judgments about object positions, viewpoints and occlusions, with two-dimensional grounding and support for three-dimensional grounding. Ask which relationship matters and which cues support the conclusion. A visual relationship should remain distinct from a calibrated physical measurement.
The native context is 262,144 tokens. The family describes books and hours-long video, with second-level event indexing and expansion to about one million tokens through configuration. Preserve timing, frame order and source identifiers when asking for an event account. Context and sampling determine the material represented in a request; the family description should not become an unconditional promise of perfect recall.
Multimodal reasoning, recognition and OCR
The Thinking edition is relevant to STEM, mathematical and logical questions that connect visual evidence with an explanation. Provide the full statement, diagram labels and units and ask for the values and relationships used before the final answer. A clear deliverable can be a delimited result, evidence table or short conclusion backed by the image and text.
The family expands visual recognition across people, products, landmarks, plants, animals and media characters. OCR covers 32 languages and targets reading under low light, blur or tilt, including specialist vocabulary and document structure. Preserve legibility and request fields or quotations that can be checked. These use cases apply documented capabilities; they do not guarantee a correct identity, transcription or mathematical conclusion for every input.
Architecture
The exact configuration identifies a vision-language mixture-of-experts model with 48 text layers, 128 experts and eight experts selected per token. Interleaved-MRoPE distributes positional information across time, width and height. DeepStack combines visual features from several encoder levels, and text–timestamp alignment helps ground events in temporal references. Use this 30B-A3B Thinking edition consistently when preparing visual reasoning tasks and comparing their results.
Preparing inputs
The family’s examples support text-only tasks, single or multiple images and video. Label visual inputs and preserve their relationship. Separate image and video pixel budgets allow a workflow to control detail; explicit image dimensions are aligned to multiples of 32. The documented video default is two frames per second, with sampling-rate or frame-count controls where supported. Keep video metadata and avoid resizing twice when the helper has already resized the media.
For a document comparison, organize the result by page identifier. For a clip, request the relevant event, interval and evidence. Reserve sequence space for thinking and the final text as well as the visual tokens.
Generation guidance and limits
The checkpoint’s generation file defaults to temperature 0.8, top_p 0.95, top_k 20 and repetition penalty 1.0. The family’s separate Thinking evaluation recipe uses temperature 0.6, presence penalty zero, repetition penalty 1.0 and a 40,960-token output allowance. These are distinct producer presets; the allowance is not an independently established hard output maximum.
The family’s extension procedure uses YaRN and suggests a factor of two or three for the interleaved positional scheme when moving from 256K to about 1M. It should not inherit another family’s factor of four. Check extracted labels, calculations and generated code against the original evidence, and keep the final artifact distinguishable from the reasoning. The practical result is a reviewable visual conclusion or code proposal with this exact Thinking edition preserved.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-vl-30b-a3b-thinking | 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 $0.00058 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0 / 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 $0.00058 / 1K tokens Output · Output tokens $0.002 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3-vl-30b-a3b-thinking",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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/qwen3-vl-30b-a3b-thinking', '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/qwen3-vl-30b-a3b-thinking', '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/qwen3-vl-30b-a3b-thinking
Use the method paths shown above.
Technical example in English
Reason about a transcribed geometry diagram\n\nUse this diagram transcription: rectangle ABCD has AB = 10 cm and BC = 6 cm; E is the midpoint of AB; segment CE is drawn. Find the area of triangle BCE. State the base and perpendicular height used, preserve the units, and give a short derivation. Do not assume any additional markings or image input.A short geometry derivation naming base BE and height BC, with the final area in square centimetres.The midpoint halves AB, and the triangle area uses one-half times the perpendicular base-height product.