Qwen3.5 highlights in the dense 27B edition
Qwen3.5-27B is a post-trained dense vision-language model. Text, images and video can participate in one task, while the result is text: an explanation, code proposal, extracted information or another requested artifact. The family’s early fusion of textual and visual training develops reasoning, coding, agent and visual abilities, with reinforcement learning across increasingly complex tasks and support for 201 languages and dialects.
That language coverage helps frame multilingual tasks but does not guarantee identical reliability for every dialect or specialist vocabulary. Provide the target language, audience and terms that should remain unchanged. For a cross-language document comparison, label each source and request a result organized by the facts or questions that need comparison.
Model overview
The exact 27B edition is dense, with 64 language-model layers and a vision encoder. Its hybrid layout alternates groups of three Gated DeltaNet layers and one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The sparse mixture-of-experts architecture described for other Qwen3.5 sizes should not be assigned to this model; its 27-billion-parameter identity is its own checkpoint specification.
The native context is 262,144 tokens, with conditional extension to 1,010,000. Qwen recommends retaining at least a 128K context where possible for complex thinking tasks, while allowing smaller settings when memory constraints require them. This is configuration guidance rather than proof that every interface exposes the full context. Source text, visual tokens, history, reasoning and the answer share the sequence allowance.
Text and visual preparation
For a technical document, keep section identifiers and ask for a requirements table or explanation supported by those sections. For a diagram, preserve labels and provide definitions that are not visible in the image. For a code task, supply the relevant function, observed behavior and required result. A useful response should identify the evidence used, separate assumptions and return the final artifact in a form that can be reviewed.
For a video question, identify the event or sequence that matters and preserve timing and frame order. The maker’s example uses two frames per second in its supported workflow. Its long-video guidance recommends a larger processor budget, corresponding to roughly 224K video tokens, for higher-rate hour-scale analysis. Those settings require configuration and do not establish a universal duration or exact event-timestamp guarantee.
Thinking and direct responses
Thinking is enabled by default and precedes the final response. enable_thinking can disable it for a direct answer. Qwen3.5 does not officially support the older /think and /nothink soft switches. The maker recommends retaining only final-answer text in historical assistant messages, rather than the thinking content. This is the Qwen3.5 contract and should remain separate from historical reasoning-preservation behavior documented for later releases.
Choose a final form before submitting the task: a short conclusion with supporting values, an extracted-field table or an explained code patch. A direct response is still a generated answer to review; a lengthy thinking transcript is not independent proof that its conclusion is correct.
Agentic usage
The maker supports tool calling and recommends Qwen-Agent with configured or MCP-connected tools, plus Qwen Code for coding-agent workflows. Give the application’s tools a clear role and return their actual observations to the model. The model can plan and revise using that feedback, while the application performs the actions. Tool support does not itself establish built-in access to the internet, files or a code-execution environment through every interface.
Ultra-long text
The extension uses YaRN with the checkpoint’s own positional configuration. A factor of four targets the approximately one-million-token workflow; a factor of two is suggested for 524,288 tokens. Static scaling can lower quality on shorter text, so the maker recommends it when the larger context is actually needed. Keep long material organized and reserve space for the required answer rather than interpret the largest context number as unlimited document recall.
Generation guidance and limits
Best Practices recommend temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0 for general thinking. 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. Direct-response reasoning uses the separate Best Practices preset of temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. Larger presence penalties can mix languages or slightly reduce quality.
Recommended generation budgets are 32,768 tokens for most tasks and 81,920 for unusually complex math or programming. They are not independent hard output ceilings. State the answer schema or delimiter explicitly and check code, calculations, quotations and visual details against the source. The useful result is a concrete text artifact with the exact dense 27B identity, inputs and response mode preserved.
", "aboutText": "Qwen3.5 highlights in the dense 27B edition\n\nQwen3.5-27B is a post-trained dense vision-language model. Text, images and video can participate in one task, while the result is text: an explanation, code proposal, extracted information or another requested artifact. The family’s early fusion of textual and visual training develops reasoning, coding, agent and visual abilities, with reinforcement learning across increasingly complex tasks and support for 201 languages and dialects.\n\nThat language coverage helps frame multilingual tasks but does not guarantee identical reliability for every dialect or specialist vocabulary. Provide the target language, audience and terms that should remain unchanged. For a cross-language document comparison, label each source and request a result organized by the facts or questions that need comparison.\n\nModel overview\n\nThe exact 27B edition is dense, with 64 language-model layers and a vision encoder. Its hybrid layout alternates groups of three Gated DeltaNet layers and one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The sparse mixture-of-experts architecture described for other Qwen3.5 sizes should not be assigned to this model; its 27-billion-parameter identity is its own checkpoint specification.\n\nThe native context is 262,144 tokens, with conditional extension to 1,010,000. Qwen recommends retaining at least a 128K context where possible for complex thinking tasks, while allowing smaller settings when memory constraints require them. This is configuration guidance rather than proof that every interface exposes the full context. Source text, visual tokens, history, reasoning and the answer share the sequence allowance.\n\nText and visual preparation\n\nFor a technical document, keep section identifiers and ask for a requirements table or explanation supported by those sections. For a diagram, preserve labels and provide definitions that are not visible in the image. For a code task, supply the relevant function, observed behavior and required result. A useful response should identify the evidence used, separate assumptions and return the final artifact in a form that can be reviewed.\n\nFor a video question, identify the event or sequence that matters and preserve timing and frame order. The maker’s example uses two frames per second in its supported workflow. Its long-video guidance recommends a larger processor budget, corresponding to roughly 224K video tokens, for higher-rate hour-scale analysis. Those settings require configuration and do not establish a universal duration or exact event-timestamp guarantee.\n\nThinking and direct responses\n\nThinking is enabled by default and precedes the final response. enable_thinking can disable it for a direct answer. Qwen3.5 does not officially support the older /think and /nothink soft switches. The maker recommends retaining only final-answer text in historical assistant messages, rather than the thinking content. This is the Qwen3.5 contract and should remain separate from historical reasoning-preservation behavior documented for later releases.\n\nChoose a final form before submitting the task: a short conclusion with supporting values, an extracted-field table or an explained code patch. A direct response is still a generated answer to review; a lengthy thinking transcript is not independent proof that its conclusion is correct.\n\nAgentic usage\n\nThe maker supports tool calling and recommends Qwen-Agent with configured or MCP-connected tools, plus Qwen Code for coding-agent workflows. Give the application’s tools a clear role and return their actual observations to the model. The model can plan and revise using that feedback, while the application performs the actions. Tool support does not itself establish built-in access to the internet, files or a code-execution environment through every interface.\n\nUltra-long text\n\nThe extension uses YaRN with the checkpoint’s own positional configuration. A factor of four targets the approximately one-million-token workflow; a factor of two is suggested for 524,288 tokens. Static scaling can lower quality on shorter text, so the maker recommends it when the larger context is actually needed. Keep long material organized and reserve space for the required answer rather than interpret the largest context number as unlimited document recall.\n\nGeneration guidance and limits\n\nBest Practices recommend temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0 for general thinking. 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. Direct-response reasoning uses the separate Best Practices preset of temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. Larger presence penalties can mix languages or slightly reduce quality.\n\nRecommended generation budgets are 32,768 tokens for most tasks and 81,920 for unusually complex math or programming. They are not independent hard output ceilings. State the answer schema or delimiter explicitly and check code, calculations, quotations and visual details against the source. The useful result is a concrete text artifact with the exact dense 27B identity, inputs and response mode preserved.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Compare a multilingual specification\n\nCompare these two product notes and write the finding in English. [EN] "Store at 5-25 C. Keep the lid closed." [FR] "Conserver entre 5 et 20 C. Garder le couvercle fermé." Identify matching instructions, the temperature discrepancy and a clarification question for the author. Preserve both source ranges without deciding which is correct.A source-labelled English comparison with matching lid guidance, differing ranges and one unresolved question.Check that 20 C and 25 C remain distinct and neither language is assumed authoritative.qwen/qwen3.5-27bQwen3.5-27B is Qwen’s dense 27-billion-parameter vision-language model for text, images and video, with textual output. It supports reasoning, coding, tool-assisted work and multilingual tasks, with selectable thinking 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.
Qwen3.5 highlights in the dense 27B edition
Qwen3.5-27B is a post-trained dense vision-language model. Text, images and video can participate in one task, while the result is text: an explanation, code proposal, extracted information or another requested artifact. The family’s early fusion of textual and visual training develops reasoning, coding, agent and visual abilities, with reinforcement learning across increasingly complex tasks and support for 201 languages and dialects.
That language coverage helps frame multilingual tasks but does not guarantee identical reliability for every dialect or specialist vocabulary. Provide the target language, audience and terms that should remain unchanged. For a cross-language document comparison, label each source and request a result organized by the facts or questions that need comparison.
Model overview
The exact 27B edition is dense, with 64 language-model layers and a vision encoder. Its hybrid layout alternates groups of three Gated DeltaNet layers and one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The sparse mixture-of-experts architecture described for other Qwen3.5 sizes should not be assigned to this model; its 27-billion-parameter identity is its own checkpoint specification.
The native context is 262,144 tokens, with conditional extension to 1,010,000. Qwen recommends retaining at least a 128K context where possible for complex thinking tasks, while allowing smaller settings when memory constraints require them. This is configuration guidance rather than proof that every interface exposes the full context. Source text, visual tokens, history, reasoning and the answer share the sequence allowance.
Text and visual preparation
For a technical document, keep section identifiers and ask for a requirements table or explanation supported by those sections. For a diagram, preserve labels and provide definitions that are not visible in the image. For a code task, supply the relevant function, observed behavior and required result. A useful response should identify the evidence used, separate assumptions and return the final artifact in a form that can be reviewed.
For a video question, identify the event or sequence that matters and preserve timing and frame order. The maker’s example uses two frames per second in its supported workflow. Its long-video guidance recommends a larger processor budget, corresponding to roughly 224K video tokens, for higher-rate hour-scale analysis. Those settings require configuration and do not establish a universal duration or exact event-timestamp guarantee.
Thinking and direct responses
Thinking is enabled by default and precedes the final response. enable_thinking can disable it for a direct answer. Qwen3.5 does not officially support the older /think and /nothink soft switches. The maker recommends retaining only final-answer text in historical assistant messages, rather than the thinking content. This is the Qwen3.5 contract and should remain separate from historical reasoning-preservation behavior documented for later releases.
Choose a final form before submitting the task: a short conclusion with supporting values, an extracted-field table or an explained code patch. A direct response is still a generated answer to review; a lengthy thinking transcript is not independent proof that its conclusion is correct.
Agentic usage
The maker supports tool calling and recommends Qwen-Agent with configured or MCP-connected tools, plus Qwen Code for coding-agent workflows. Give the application’s tools a clear role and return their actual observations to the model. The model can plan and revise using that feedback, while the application performs the actions. Tool support does not itself establish built-in access to the internet, files or a code-execution environment through every interface.
Ultra-long text
The extension uses YaRN with the checkpoint’s own positional configuration. A factor of four targets the approximately one-million-token workflow; a factor of two is suggested for 524,288 tokens. Static scaling can lower quality on shorter text, so the maker recommends it when the larger context is actually needed. Keep long material organized and reserve space for the required answer rather than interpret the largest context number as unlimited document recall.
Generation guidance and limits
Best Practices recommend temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0 for general thinking. 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. Direct-response reasoning uses the separate Best Practices preset of temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. Larger presence penalties can mix languages or slightly reduce quality.
Recommended generation budgets are 32,768 tokens for most tasks and 81,920 for unusually complex math or programming. They are not independent hard output ceilings. State the answer schema or delimiter explicitly and check code, calculations, quotations and visual details against the source. The useful result is a concrete text artifact with the exact dense 27B identity, inputs and response mode preserved.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.5-27b | 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 $0.00052 / 1K tokens Output · Output tokens $0.00412 / 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 $0.00052 / 1K tokens Output · Output tokens $0.00412 / 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/qwen3.5-27b", "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/qwen3.5-27b',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/qwen3.5-27b',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/qwen3.5-27b
Use the method paths shown above.
qwen/qwen3.5-27bQwen3.5-27B is Qwen’s dense 27-billion-parameter vision-language model for text, images and video, with textual output. It supports reasoning, coding, tool-assisted work and multilingual tasks, with selectable thinking 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.
Qwen3.5 highlights in the dense 27B edition
Qwen3.5-27B is a post-trained dense vision-language model. Text, images and video can participate in one task, while the result is text: an explanation, code proposal, extracted information or another requested artifact. The family’s early fusion of textual and visual training develops reasoning, coding, agent and visual abilities, with reinforcement learning across increasingly complex tasks and support for 201 languages and dialects.
That language coverage helps frame multilingual tasks but does not guarantee identical reliability for every dialect or specialist vocabulary. Provide the target language, audience and terms that should remain unchanged. For a cross-language document comparison, label each source and request a result organized by the facts or questions that need comparison.
Model overview
The exact 27B edition is dense, with 64 language-model layers and a vision encoder. Its hybrid layout alternates groups of three Gated DeltaNet layers and one gated-attention layer, using dense feed-forward networks. Multi-token prediction is included in training. The sparse mixture-of-experts architecture described for other Qwen3.5 sizes should not be assigned to this model; its 27-billion-parameter identity is its own checkpoint specification.
The native context is 262,144 tokens, with conditional extension to 1,010,000. Qwen recommends retaining at least a 128K context where possible for complex thinking tasks, while allowing smaller settings when memory constraints require them. This is configuration guidance rather than proof that every interface exposes the full context. Source text, visual tokens, history, reasoning and the answer share the sequence allowance.
Text and visual preparation
For a technical document, keep section identifiers and ask for a requirements table or explanation supported by those sections. For a diagram, preserve labels and provide definitions that are not visible in the image. For a code task, supply the relevant function, observed behavior and required result. A useful response should identify the evidence used, separate assumptions and return the final artifact in a form that can be reviewed.
For a video question, identify the event or sequence that matters and preserve timing and frame order. The maker’s example uses two frames per second in its supported workflow. Its long-video guidance recommends a larger processor budget, corresponding to roughly 224K video tokens, for higher-rate hour-scale analysis. Those settings require configuration and do not establish a universal duration or exact event-timestamp guarantee.
Thinking and direct responses
Thinking is enabled by default and precedes the final response. enable_thinking can disable it for a direct answer. Qwen3.5 does not officially support the older /think and /nothink soft switches. The maker recommends retaining only final-answer text in historical assistant messages, rather than the thinking content. This is the Qwen3.5 contract and should remain separate from historical reasoning-preservation behavior documented for later releases.
Choose a final form before submitting the task: a short conclusion with supporting values, an extracted-field table or an explained code patch. A direct response is still a generated answer to review; a lengthy thinking transcript is not independent proof that its conclusion is correct.
Agentic usage
The maker supports tool calling and recommends Qwen-Agent with configured or MCP-connected tools, plus Qwen Code for coding-agent workflows. Give the application’s tools a clear role and return their actual observations to the model. The model can plan and revise using that feedback, while the application performs the actions. Tool support does not itself establish built-in access to the internet, files or a code-execution environment through every interface.
Ultra-long text
The extension uses YaRN with the checkpoint’s own positional configuration. A factor of four targets the approximately one-million-token workflow; a factor of two is suggested for 524,288 tokens. Static scaling can lower quality on shorter text, so the maker recommends it when the larger context is actually needed. Keep long material organized and reserve space for the required answer rather than interpret the largest context number as unlimited document recall.
Generation guidance and limits
Best Practices recommend temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 1.5 and repetition penalty 1.0 for general thinking. 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. Direct-response reasoning uses the separate Best Practices preset of temperature 1.0, top_p 1.0, top_k 40 and presence penalty 2.0. Larger presence penalties can mix languages or slightly reduce quality.
Recommended generation budgets are 32,768 tokens for most tasks and 81,920 for unusually complex math or programming. They are not independent hard output ceilings. State the answer schema or delimiter explicitly and check code, calculations, quotations and visual details against the source. The useful result is a concrete text artifact with the exact dense 27B identity, inputs and response mode preserved.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.5-27b | 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 $0.00052 / 1K tokens Output · Output tokens $0.00412 / 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 $0.00052 / 1K tokens Output · Output tokens $0.00412 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3.5-27b",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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/qwen3.5-27b', '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/qwen3.5-27b', '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/qwen3.5-27b
Use the method paths shown above.
Technical example in English
Compare a multilingual specification\n\nCompare these two product notes and write the finding in English. [EN] "Store at 5-25 C. Keep the lid closed." [FR] "Conserver entre 5 et 20 C. Garder le couvercle fermé." Identify matching instructions, the temperature discrepancy and a clarification question for the author. Preserve both source ranges without deciding which is correct.A source-labelled English comparison with matching lid guidance, differing ranges and one unresolved question.Check that 20 C and 25 C remain distinct and neither language is assumed authoritative.