Qwen3.5 highlights in the 9B dense edition
Qwen3.5-9B is the post-trained 9-billion-parameter dense vision-language checkpoint. It accepts text, images and video and returns text. The family brings together reasoning, coding, agents and visual understanding through early fusion of text and visual training, reinforcement learning on progressively complex tasks and support for 201 languages and dialects. That language coverage is a family capability, not a guarantee of identical reliability for every language or subject.
For multilingual document work, state the target language, preserve technical terms and ask the result to identify the source fields or passages it uses. For a visual question, provide a readable image and a precise task. For a small coding request, supply the input, desired behavior and relevant code. The model does not generate images or videos when analyzing visual material.
Model overview
This exact 9B model is dense, with 32 language-model layers and a vision encoder. Its hybrid layout repeats three Gated DeltaNet layers followed by one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The family also contains sparse expert models, but their total-versus-activated parameter descriptions should not be applied to this dense edition.
The native context is 262,144 tokens, with conditional extension to 1,010,000. The model’s configuration and numerical size should remain distinct from its Base checkpoint, other Qwen3.5 sizes and quantized builds. A smaller parameter count identifies the edition; it does not by itself establish a universal response time or quality ranking. For complex thinking tasks, Qwen recommends retaining a configured context of at least 128K tokens where possible.
Inputs and response modes
The maker shows text-only, image and video questions. Give a document question with the relevant page and expected fields, or a video question with the event and time range that matter. Preserve labels, units and frame order. Ask for a result that can be compared with the source, such as an extracted-value table or an explanation that cites the visible evidence.
Thinking is enabled by default. enable_thinking can disable reasoning for a direct response, while the older /think and /nothink soft prompt switches are not officially supported. Use the maker’s template and keep reasoning distinct from final text when the interface provides both. The Qwen3.5 history recommendation is to retain final answers without their thinking content; newer historical-preservation behavior should not be imported into this release.
Agentic usage
Qwen documents tool calling and recommends Qwen-Agent, including configured or MCP-connected tools, and Qwen Code for coding-agent work. State the permitted tools and the observable result. A connected tool can provide a calculation, source retrieval or real execution feedback; the model can then use that information in its answer. The tools and external actions belong to the application, so the model name alone does not establish access or execution.
Ultra-long text
The documented extension uses YaRN with this checkpoint’s own positional settings. A factor of four targets the roughly one-million-token workflow; for 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer context. Input, visual representation, history, reasoning and final output share the sequence budget. Organize long material by document or code identifier and leave room for the requested artifact.
Generation recommendations
The card’s Best Practices distinguish several presets. General thinking uses temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0. Precise coding uses temperature 0.6 and presence penalty zero. General direct responses use temperature 0.7, top_p 0.8 and presence penalty 1.5; top_k is 20, min_p is 0 and repetition penalty is 1.0. For direct-response reasoning tasks, Best Practices instead recommend temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. These are task-specific recipes, not values to mix indiscriminately. Larger presence penalties can cause language mixing or a slight quality decrease.
Qwen recommends a 32,768-token generation allowance for most queries and 81,920 for unusually complex math or programming questions. Those are recommendations rather than independent hard output ceilings. Specify the answer format, such as a delimited mathematical result or an answer field, and verify the conclusion against the source.
Video detail and substantive limits
The maker’s video example defaults to two frames per second in its supported workflow. Its hour-scale guidance expands the video processor’s budget to roughly 224K video tokens for higher-rate sampling. These are configuration-dependent examples, not a universal duration or event-timing guarantee. Keep small writing and important scene transitions readable, and check extracted text, calculations and generated code before using the result. The practical value of the 9B edition is its documented joint text/visual task scope and controllable response mode, with its exact dense identity preserved.
", "aboutText": "Qwen3.5 highlights in the 9B dense edition\n\nQwen3.5-9B is the post-trained 9-billion-parameter dense vision-language checkpoint. It accepts text, images and video and returns text. The family brings together reasoning, coding, agents and visual understanding through early fusion of text and visual training, reinforcement learning on progressively complex tasks and support for 201 languages and dialects. That language coverage is a family capability, not a guarantee of identical reliability for every language or subject.\n\nFor multilingual document work, state the target language, preserve technical terms and ask the result to identify the source fields or passages it uses. For a visual question, provide a readable image and a precise task. For a small coding request, supply the input, desired behavior and relevant code. The model does not generate images or videos when analyzing visual material.\n\nModel overview\n\nThis exact 9B model is dense, with 32 language-model layers and a vision encoder. Its hybrid layout repeats three Gated DeltaNet layers followed by one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The family also contains sparse expert models, but their total-versus-activated parameter descriptions should not be applied to this dense edition.\n\nThe native context is 262,144 tokens, with conditional extension to 1,010,000. The model’s configuration and numerical size should remain distinct from its Base checkpoint, other Qwen3.5 sizes and quantized builds. A smaller parameter count identifies the edition; it does not by itself establish a universal response time or quality ranking. For complex thinking tasks, Qwen recommends retaining a configured context of at least 128K tokens where possible.\n\nInputs and response modes\n\nThe maker shows text-only, image and video questions. Give a document question with the relevant page and expected fields, or a video question with the event and time range that matter. Preserve labels, units and frame order. Ask for a result that can be compared with the source, such as an extracted-value table or an explanation that cites the visible evidence.\n\nThinking is enabled by default. enable_thinking can disable reasoning for a direct response, while the older /think and /nothink soft prompt switches are not officially supported. Use the maker’s template and keep reasoning distinct from final text when the interface provides both. The Qwen3.5 history recommendation is to retain final answers without their thinking content; newer historical-preservation behavior should not be imported into this release.\n\nAgentic usage\n\nQwen documents tool calling and recommends Qwen-Agent, including configured or MCP-connected tools, and Qwen Code for coding-agent work. State the permitted tools and the observable result. A connected tool can provide a calculation, source retrieval or real execution feedback; the model can then use that information in its answer. The tools and external actions belong to the application, so the model name alone does not establish access or execution.\n\nUltra-long text\n\nThe documented extension uses YaRN with this checkpoint’s own positional settings. A factor of four targets the roughly one-million-token workflow; for 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer context. Input, visual representation, history, reasoning and final output share the sequence budget. Organize long material by document or code identifier and leave room for the requested artifact.\n\nGeneration recommendations\n\nThe card’s Best Practices distinguish several presets. General thinking uses temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0. Precise coding uses temperature 0.6 and presence penalty zero. General direct responses use temperature 0.7, top_p 0.8 and presence penalty 1.5; top_k is 20, min_p is 0 and repetition penalty is 1.0. For direct-response reasoning tasks, Best Practices instead recommend temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. These are task-specific recipes, not values to mix indiscriminately. Larger presence penalties can cause language mixing or a slight quality decrease.\n\nQwen recommends a 32,768-token generation allowance for most queries and 81,920 for unusually complex math or programming questions. Those are recommendations rather than independent hard output ceilings. Specify the answer format, such as a delimited mathematical result or an answer field, and verify the conclusion against the source.\n\nVideo detail and substantive limits\n\nThe maker’s video example defaults to two frames per second in its supported workflow. Its hour-scale guidance expands the video processor’s budget to roughly 224K video tokens for higher-rate sampling. These are configuration-dependent examples, not a universal duration or event-timing guarantee. Keep small writing and important scene transitions readable, and check extracted text, calculations and generated code before using the result. The practical value of the 9B edition is its documented joint text/visual task scope and controllable response mode, with its exact dense identity preserved.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Implement a small text normaliser\n\nWrite a Python function normalise_code(value) for product codes. Trim leading and trailing spaces, convert letters to uppercase and preserve internal hyphens and digits. Empty or whitespace-only input returns an empty string. Provide tests for ' ab-17 ', 'X-0' and ' '. Do not remove punctuation beyond surrounding whitespace or use external packages.The complete function and three assert tests preserving the specified code semantics.The function converts ' ab-17 ' to 'AB-17' while retaining the hyphen and digit order.qwen/qwen3.5-9bQwen3.5-9B is Qwen’s dense 9-billion-parameter vision-language model for text, images and video, returning text. It combines reasoning, coding, tool calling and multilingual work, with selectable thinking and a native 262,144-token context that can be extended through configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3.5 highlights in the 9B dense edition
Qwen3.5-9B is the post-trained 9-billion-parameter dense vision-language checkpoint. It accepts text, images and video and returns text. The family brings together reasoning, coding, agents and visual understanding through early fusion of text and visual training, reinforcement learning on progressively complex tasks and support for 201 languages and dialects. That language coverage is a family capability, not a guarantee of identical reliability for every language or subject.
For multilingual document work, state the target language, preserve technical terms and ask the result to identify the source fields or passages it uses. For a visual question, provide a readable image and a precise task. For a small coding request, supply the input, desired behavior and relevant code. The model does not generate images or videos when analyzing visual material.
Model overview
This exact 9B model is dense, with 32 language-model layers and a vision encoder. Its hybrid layout repeats three Gated DeltaNet layers followed by one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The family also contains sparse expert models, but their total-versus-activated parameter descriptions should not be applied to this dense edition.
The native context is 262,144 tokens, with conditional extension to 1,010,000. The model’s configuration and numerical size should remain distinct from its Base checkpoint, other Qwen3.5 sizes and quantized builds. A smaller parameter count identifies the edition; it does not by itself establish a universal response time or quality ranking. For complex thinking tasks, Qwen recommends retaining a configured context of at least 128K tokens where possible.
Inputs and response modes
The maker shows text-only, image and video questions. Give a document question with the relevant page and expected fields, or a video question with the event and time range that matter. Preserve labels, units and frame order. Ask for a result that can be compared with the source, such as an extracted-value table or an explanation that cites the visible evidence.
Thinking is enabled by default. enable_thinking can disable reasoning for a direct response, while the older /think and /nothink soft prompt switches are not officially supported. Use the maker’s template and keep reasoning distinct from final text when the interface provides both. The Qwen3.5 history recommendation is to retain final answers without their thinking content; newer historical-preservation behavior should not be imported into this release.
Agentic usage
Qwen documents tool calling and recommends Qwen-Agent, including configured or MCP-connected tools, and Qwen Code for coding-agent work. State the permitted tools and the observable result. A connected tool can provide a calculation, source retrieval or real execution feedback; the model can then use that information in its answer. The tools and external actions belong to the application, so the model name alone does not establish access or execution.
Ultra-long text
The documented extension uses YaRN with this checkpoint’s own positional settings. A factor of four targets the roughly one-million-token workflow; for 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer context. Input, visual representation, history, reasoning and final output share the sequence budget. Organize long material by document or code identifier and leave room for the requested artifact.
Generation recommendations
The card’s Best Practices distinguish several presets. General thinking uses temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0. Precise coding uses temperature 0.6 and presence penalty zero. General direct responses use temperature 0.7, top_p 0.8 and presence penalty 1.5; top_k is 20, min_p is 0 and repetition penalty is 1.0. For direct-response reasoning tasks, Best Practices instead recommend temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. These are task-specific recipes, not values to mix indiscriminately. Larger presence penalties can cause language mixing or a slight quality decrease.
Qwen recommends a 32,768-token generation allowance for most queries and 81,920 for unusually complex math or programming questions. Those are recommendations rather than independent hard output ceilings. Specify the answer format, such as a delimited mathematical result or an answer field, and verify the conclusion against the source.
Video detail and substantive limits
The maker’s video example defaults to two frames per second in its supported workflow. Its hour-scale guidance expands the video processor’s budget to roughly 224K video tokens for higher-rate sampling. These are configuration-dependent examples, not a universal duration or event-timing guarantee. Keep small writing and important scene transitions readable, and check extracted text, calculations and generated code before using the result. The practical value of the 9B edition is its documented joint text/visual task scope and controllable response mode, with its exact dense identity preserved.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.5-9b | 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 $0.0002 / 1K tokens Output · Output tokens $0.0003 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 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 $0.0002 / 1K tokens Output · Output tokens $0.0003 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "qwen/qwen3.5-9b", "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/qwen3.5-9b',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/qwen3.5-9b',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/qwen3.5-9b
Use the method paths shown above.
qwen/qwen3.5-9bQwen3.5-9B is Qwen’s dense 9-billion-parameter vision-language model for text, images and video, returning text. It combines reasoning, coding, tool calling and multilingual work, with selectable thinking and a native 262,144-token context that can be extended through configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3.5 highlights in the 9B dense edition
Qwen3.5-9B is the post-trained 9-billion-parameter dense vision-language checkpoint. It accepts text, images and video and returns text. The family brings together reasoning, coding, agents and visual understanding through early fusion of text and visual training, reinforcement learning on progressively complex tasks and support for 201 languages and dialects. That language coverage is a family capability, not a guarantee of identical reliability for every language or subject.
For multilingual document work, state the target language, preserve technical terms and ask the result to identify the source fields or passages it uses. For a visual question, provide a readable image and a precise task. For a small coding request, supply the input, desired behavior and relevant code. The model does not generate images or videos when analyzing visual material.
Model overview
This exact 9B model is dense, with 32 language-model layers and a vision encoder. Its hybrid layout repeats three Gated DeltaNet layers followed by one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The family also contains sparse expert models, but their total-versus-activated parameter descriptions should not be applied to this dense edition.
The native context is 262,144 tokens, with conditional extension to 1,010,000. The model’s configuration and numerical size should remain distinct from its Base checkpoint, other Qwen3.5 sizes and quantized builds. A smaller parameter count identifies the edition; it does not by itself establish a universal response time or quality ranking. For complex thinking tasks, Qwen recommends retaining a configured context of at least 128K tokens where possible.
Inputs and response modes
The maker shows text-only, image and video questions. Give a document question with the relevant page and expected fields, or a video question with the event and time range that matter. Preserve labels, units and frame order. Ask for a result that can be compared with the source, such as an extracted-value table or an explanation that cites the visible evidence.
Thinking is enabled by default. enable_thinking can disable reasoning for a direct response, while the older /think and /nothink soft prompt switches are not officially supported. Use the maker’s template and keep reasoning distinct from final text when the interface provides both. The Qwen3.5 history recommendation is to retain final answers without their thinking content; newer historical-preservation behavior should not be imported into this release.
Agentic usage
Qwen documents tool calling and recommends Qwen-Agent, including configured or MCP-connected tools, and Qwen Code for coding-agent work. State the permitted tools and the observable result. A connected tool can provide a calculation, source retrieval or real execution feedback; the model can then use that information in its answer. The tools and external actions belong to the application, so the model name alone does not establish access or execution.
Ultra-long text
The documented extension uses YaRN with this checkpoint’s own positional settings. A factor of four targets the roughly one-million-token workflow; for 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer context. Input, visual representation, history, reasoning and final output share the sequence budget. Organize long material by document or code identifier and leave room for the requested artifact.
Generation recommendations
The card’s Best Practices distinguish several presets. General thinking uses temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0. Precise coding uses temperature 0.6 and presence penalty zero. General direct responses use temperature 0.7, top_p 0.8 and presence penalty 1.5; top_k is 20, min_p is 0 and repetition penalty is 1.0. For direct-response reasoning tasks, Best Practices instead recommend temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. These are task-specific recipes, not values to mix indiscriminately. Larger presence penalties can cause language mixing or a slight quality decrease.
Qwen recommends a 32,768-token generation allowance for most queries and 81,920 for unusually complex math or programming questions. Those are recommendations rather than independent hard output ceilings. Specify the answer format, such as a delimited mathematical result or an answer field, and verify the conclusion against the source.
Video detail and substantive limits
The maker’s video example defaults to two frames per second in its supported workflow. Its hour-scale guidance expands the video processor’s budget to roughly 224K video tokens for higher-rate sampling. These are configuration-dependent examples, not a universal duration or event-timing guarantee. Keep small writing and important scene transitions readable, and check extracted text, calculations and generated code before using the result. The practical value of the 9B edition is its documented joint text/visual task scope and controllable response mode, with its exact dense identity preserved.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.5-9b | 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 $0.0002 / 1K tokens Output · Output tokens $0.0003 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 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 $0.0002 / 1K tokens Output · Output tokens $0.0003 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3.5-9b",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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/qwen3.5-9b', '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/qwen3.5-9b', '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/qwen3.5-9b
Use the method paths shown above.
Technical example in English
Implement a small text normaliser\n\nWrite a Python function normalise_code(value) for product codes. Trim leading and trailing spaces, convert letters to uppercase and preserve internal hyphens and digits. Empty or whitespace-only input returns an empty string. Provide tests for ' ab-17 ', 'X-0' and ' '. Do not remove punctuation beyond surrounding whitespace or use external packages.The complete function and three assert tests preserving the specified code semantics.The function converts ' ab-17 ' to 'AB-17' while retaining the hyphen and digit order.