Qwen3.8 highlights in a dense vision-language model
Qwen3.8-27B brings Qwen3.8’s coding, professional-work, research and agent-task improvements into a 27-billion-parameter dense model. It accepts text, images and video and returns text. Its documented visual scope includes STEM diagrams, document pages and hour-scale video. It is a distinct checkpoint from the much larger, text-only Qwen3.8-2.4T-A95B, so their input capabilities and thinking contracts should remain separate.
Qwen emphasizes autonomous planning, the use of environment feedback and integration with development harnesses. For an agent task, supply the goal, relevant material, allowed tools and evidence that would show completion. A model can propose and revise actions using returned results; connected software performs those actions. Ask for the final artifact and unresolved conditions, rather than treat a plan as an executed result.
Model overview
The 27B model has 64 language-model layers. Its hybrid layout groups three Gated DeltaNet layers followed by one gated-attention layer, with dense feed-forward networks. It is trained with multi-token prediction and includes a vision encoder. These architectural choices describe the released model; they do not imply that every deployment has identical memory use or latency.
The native context is 262,144 tokens, with documented extension to 1,000,000. Text, visual representation, conversation history, thinking and final output share the configured sequence budget. The maker separately discusses a managed edition with additional features; those features should not automatically be assigned to the open checkpoint.
Using text, images and video
For a code task, provide the relevant function or module, a failing input and expected behavior, then request a patch with the affected assumptions and checks. For a diagram, preserve labels and give the complete question. For a document, ask for named fields or conclusions with page references. For video, state the event of interest and preserve its timing.
The maker’s video example uses two frames per second by default in its supported serving workflow, with optional sampling controls. A broad scene summary and a brief event may need different sampling. The long-video guidance recommends expanding the video processor’s budget to support higher-rate sampling for hour-scale material, with an example corresponding to roughly 224K video tokens. This requires configuration and competes with other context content; it is not a universal clip-duration guarantee.
Thinking and historical reasoning
Thinking is enabled by default and can be disabled per request through the maker’s enable_thinking control. reasoning_effort supports xhigh, medium and low, with xhigh the default. Lower effort can shorten an individual turn, but Qwen warns that insufficient analysis in agent tasks can cause failures and repeated work, increasing total completion time.
preserve_thinking is enabled by default, retaining reasoning blocks from historical messages for continuity. Disabling it retains only the thinking blocks generated while handling the latest user message. This is different from older Qwen guidance that discards historical reasoning entirely. Keep reasoning and final content distinct in the interface, and preserve the exact conversation format expected by the checkpoint. Maker controls do not establish which parameters another application exposes.
Generation guidance
For thinking, Qwen recommends temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 0 and repetition penalty 1.0. For direct responses, it recommends 0.7, 0.8, 20, 0, 1.5 and 1.0 respectively. A higher presence penalty can reduce repetition but may mix languages or slightly reduce quality.
Within a configured one-million-token workflow that supports separate limits, the maker recommends allowances of 262,144 tokens for reasoning and 131,072 for the final response. These are conditional budget recommendations, not independent hard output ceilings or allocations that fit simultaneously inside the native context.
Ultra-long context and limits
The documented extension uses YaRN with the model’s own interleaved positional settings. A factor of four targets the one-million-token workflow; for a workload around 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text quality, so Qwen advises enabling it when the longer context is needed. Do not copy a different Qwen-VL edition’s extrapolation factor without its own recipe.
Keep long input organized by document, page or code identifier and request a checkable result. Verify calculations, small visual labels and generated patches against the source. Long context, retained reasoning and larger video budgets help carry more evidence through a task; they do not by themselves establish complete recall, correct reasoning or a completed external action.
", "aboutText": "Qwen3.8 highlights in a dense vision-language model\n\nQwen3.8-27B brings Qwen3.8’s coding, professional-work, research and agent-task improvements into a 27-billion-parameter dense model. It accepts text, images and video and returns text. Its documented visual scope includes STEM diagrams, document pages and hour-scale video. It is a distinct checkpoint from the much larger, text-only Qwen3.8-2.4T-A95B, so their input capabilities and thinking contracts should remain separate.\n\nQwen emphasizes autonomous planning, the use of environment feedback and integration with development harnesses. For an agent task, supply the goal, relevant material, allowed tools and evidence that would show completion. A model can propose and revise actions using returned results; connected software performs those actions. Ask for the final artifact and unresolved conditions, rather than treat a plan as an executed result.\n\nModel overview\n\nThe 27B model has 64 language-model layers. Its hybrid layout groups three Gated DeltaNet layers followed by one gated-attention layer, with dense feed-forward networks. It is trained with multi-token prediction and includes a vision encoder. These architectural choices describe the released model; they do not imply that every deployment has identical memory use or latency.\n\nThe native context is 262,144 tokens, with documented extension to 1,000,000. Text, visual representation, conversation history, thinking and final output share the configured sequence budget. The maker separately discusses a managed edition with additional features; those features should not automatically be assigned to the open checkpoint.\n\nUsing text, images and video\n\nFor a code task, provide the relevant function or module, a failing input and expected behavior, then request a patch with the affected assumptions and checks. For a diagram, preserve labels and give the complete question. For a document, ask for named fields or conclusions with page references. For video, state the event of interest and preserve its timing.\n\nThe maker’s video example uses two frames per second by default in its supported serving workflow, with optional sampling controls. A broad scene summary and a brief event may need different sampling. The long-video guidance recommends expanding the video processor’s budget to support higher-rate sampling for hour-scale material, with an example corresponding to roughly 224K video tokens. This requires configuration and competes with other context content; it is not a universal clip-duration guarantee.\n\nThinking and historical reasoning\n\nThinking is enabled by default and can be disabled per request through the maker’s enable_thinking control. reasoning_effort supports xhigh, medium and low, with xhigh the default. Lower effort can shorten an individual turn, but Qwen warns that insufficient analysis in agent tasks can cause failures and repeated work, increasing total completion time.\n\npreserve_thinking is enabled by default, retaining reasoning blocks from historical messages for continuity. Disabling it retains only the thinking blocks generated while handling the latest user message. This is different from older Qwen guidance that discards historical reasoning entirely. Keep reasoning and final content distinct in the interface, and preserve the exact conversation format expected by the checkpoint. Maker controls do not establish which parameters another application exposes.\n\nGeneration guidance\n\nFor thinking, Qwen recommends temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 0 and repetition penalty 1.0. For direct responses, it recommends 0.7, 0.8, 20, 0, 1.5 and 1.0 respectively. A higher presence penalty can reduce repetition but may mix languages or slightly reduce quality.\n\nWithin a configured one-million-token workflow that supports separate limits, the maker recommends allowances of 262,144 tokens for reasoning and 131,072 for the final response. These are conditional budget recommendations, not independent hard output ceilings or allocations that fit simultaneously inside the native context.\n\nUltra-long context and limits\n\nThe documented extension uses YaRN with the model’s own interleaved positional settings. A factor of four targets the one-million-token workflow; for a workload around 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text quality, so Qwen advises enabling it when the longer context is needed. Do not copy a different Qwen-VL edition’s extrapolation factor without its own recipe.\n\nKeep long input organized by document, page or code identifier and request a checkable result. Verify calculations, small visual labels and generated patches against the source. Long context, retained reasoning and larger video budgets help carry more evidence through a task; they do not by themselves establish complete recall, correct reasoning or a completed external action.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Analyze a circuit described in text\n\nA written circuit description gives two 6-ohm resistors connected in parallel across an ideal 12-volt source. Assume ideal wires and constant resistance. Derive the equivalent resistance, total current and current through each branch. Show the formulas and verify that branch currents add to the total. No circuit image or execution tool is provided.A short derivation with units, three requested quantities and a current-sum check.Equivalent resistance is 3 ohms and the two equal branch currents sum to the derived total current.qwen/qwen3.8-27bQwen3.8-27B is Qwen’s dense vision-language model for text, images and video, returning text. It combines coding and long-step agent work with selectable thinking, reasoning preservation and a native 262,144-token context, extendable to one million with configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3.8 highlights in a dense vision-language model
Qwen3.8-27B brings Qwen3.8’s coding, professional-work, research and agent-task improvements into a 27-billion-parameter dense model. It accepts text, images and video and returns text. Its documented visual scope includes STEM diagrams, document pages and hour-scale video. It is a distinct checkpoint from the much larger, text-only Qwen3.8-2.4T-A95B, so their input capabilities and thinking contracts should remain separate.
Qwen emphasizes autonomous planning, the use of environment feedback and integration with development harnesses. For an agent task, supply the goal, relevant material, allowed tools and evidence that would show completion. A model can propose and revise actions using returned results; connected software performs those actions. Ask for the final artifact and unresolved conditions, rather than treat a plan as an executed result.
Model overview
The 27B model has 64 language-model layers. Its hybrid layout groups three Gated DeltaNet layers followed by one gated-attention layer, with dense feed-forward networks. It is trained with multi-token prediction and includes a vision encoder. These architectural choices describe the released model; they do not imply that every deployment has identical memory use or latency.
The native context is 262,144 tokens, with documented extension to 1,000,000. Text, visual representation, conversation history, thinking and final output share the configured sequence budget. The maker separately discusses a managed edition with additional features; those features should not automatically be assigned to the open checkpoint.
Using text, images and video
For a code task, provide the relevant function or module, a failing input and expected behavior, then request a patch with the affected assumptions and checks. For a diagram, preserve labels and give the complete question. For a document, ask for named fields or conclusions with page references. For video, state the event of interest and preserve its timing.
The maker’s video example uses two frames per second by default in its supported serving workflow, with optional sampling controls. A broad scene summary and a brief event may need different sampling. The long-video guidance recommends expanding the video processor’s budget to support higher-rate sampling for hour-scale material, with an example corresponding to roughly 224K video tokens. This requires configuration and competes with other context content; it is not a universal clip-duration guarantee.
Thinking and historical reasoning
Thinking is enabled by default and can be disabled per request through the maker’s enable_thinking control. reasoning_effort supports xhigh, medium and low, with xhigh the default. Lower effort can shorten an individual turn, but Qwen warns that insufficient analysis in agent tasks can cause failures and repeated work, increasing total completion time.
preserve_thinking is enabled by default, retaining reasoning blocks from historical messages for continuity. Disabling it retains only the thinking blocks generated while handling the latest user message. This is different from older Qwen guidance that discards historical reasoning entirely. Keep reasoning and final content distinct in the interface, and preserve the exact conversation format expected by the checkpoint. Maker controls do not establish which parameters another application exposes.
Generation guidance
For thinking, Qwen recommends temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 0 and repetition penalty 1.0. For direct responses, it recommends 0.7, 0.8, 20, 0, 1.5 and 1.0 respectively. A higher presence penalty can reduce repetition but may mix languages or slightly reduce quality.
Within a configured one-million-token workflow that supports separate limits, the maker recommends allowances of 262,144 tokens for reasoning and 131,072 for the final response. These are conditional budget recommendations, not independent hard output ceilings or allocations that fit simultaneously inside the native context.
Ultra-long context and limits
The documented extension uses YaRN with the model’s own interleaved positional settings. A factor of four targets the one-million-token workflow; for a workload around 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text quality, so Qwen advises enabling it when the longer context is needed. Do not copy a different Qwen-VL edition’s extrapolation factor without its own recipe.
Keep long input organized by document, page or code identifier and request a checkable result. Verify calculations, small visual labels and generated patches against the source. Long context, retained reasoning and larger video budgets help carry more evidence through a task; they do not by themselves establish complete recall, correct reasoning or a completed external action.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.8-27b | 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 $0.00085 / 1K tokens Output · Output tokens $0.0051 / 1K tokens Input · Cache read $0.00017 / 1K tokens Input · Cache write $0.0010625 / 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 $0.00085 / 1K tokens Output · Output tokens $0.0051 / 1K tokens Input · Cache read $0.00017 / 1K tokens Input · Cache write $0.0010625 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "qwen/qwen3.8-27b", "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/qwen3.8-27b',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/qwen3.8-27b',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/qwen3.8-27b
Use the method paths shown above.
qwen/qwen3.8-27bQwen3.8-27B is Qwen’s dense vision-language model for text, images and video, returning text. It combines coding and long-step agent work with selectable thinking, reasoning preservation and a native 262,144-token context, extendable to one million with configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3.8 highlights in a dense vision-language model
Qwen3.8-27B brings Qwen3.8’s coding, professional-work, research and agent-task improvements into a 27-billion-parameter dense model. It accepts text, images and video and returns text. Its documented visual scope includes STEM diagrams, document pages and hour-scale video. It is a distinct checkpoint from the much larger, text-only Qwen3.8-2.4T-A95B, so their input capabilities and thinking contracts should remain separate.
Qwen emphasizes autonomous planning, the use of environment feedback and integration with development harnesses. For an agent task, supply the goal, relevant material, allowed tools and evidence that would show completion. A model can propose and revise actions using returned results; connected software performs those actions. Ask for the final artifact and unresolved conditions, rather than treat a plan as an executed result.
Model overview
The 27B model has 64 language-model layers. Its hybrid layout groups three Gated DeltaNet layers followed by one gated-attention layer, with dense feed-forward networks. It is trained with multi-token prediction and includes a vision encoder. These architectural choices describe the released model; they do not imply that every deployment has identical memory use or latency.
The native context is 262,144 tokens, with documented extension to 1,000,000. Text, visual representation, conversation history, thinking and final output share the configured sequence budget. The maker separately discusses a managed edition with additional features; those features should not automatically be assigned to the open checkpoint.
Using text, images and video
For a code task, provide the relevant function or module, a failing input and expected behavior, then request a patch with the affected assumptions and checks. For a diagram, preserve labels and give the complete question. For a document, ask for named fields or conclusions with page references. For video, state the event of interest and preserve its timing.
The maker’s video example uses two frames per second by default in its supported serving workflow, with optional sampling controls. A broad scene summary and a brief event may need different sampling. The long-video guidance recommends expanding the video processor’s budget to support higher-rate sampling for hour-scale material, with an example corresponding to roughly 224K video tokens. This requires configuration and competes with other context content; it is not a universal clip-duration guarantee.
Thinking and historical reasoning
Thinking is enabled by default and can be disabled per request through the maker’s enable_thinking control. reasoning_effort supports xhigh, medium and low, with xhigh the default. Lower effort can shorten an individual turn, but Qwen warns that insufficient analysis in agent tasks can cause failures and repeated work, increasing total completion time.
preserve_thinking is enabled by default, retaining reasoning blocks from historical messages for continuity. Disabling it retains only the thinking blocks generated while handling the latest user message. This is different from older Qwen guidance that discards historical reasoning entirely. Keep reasoning and final content distinct in the interface, and preserve the exact conversation format expected by the checkpoint. Maker controls do not establish which parameters another application exposes.
Generation guidance
For thinking, Qwen recommends temperature 1.0, top_p 0.95, top_k 20, min_p 0, presence penalty 0 and repetition penalty 1.0. For direct responses, it recommends 0.7, 0.8, 20, 0, 1.5 and 1.0 respectively. A higher presence penalty can reduce repetition but may mix languages or slightly reduce quality.
Within a configured one-million-token workflow that supports separate limits, the maker recommends allowances of 262,144 tokens for reasoning and 131,072 for the final response. These are conditional budget recommendations, not independent hard output ceilings or allocations that fit simultaneously inside the native context.
Ultra-long context and limits
The documented extension uses YaRN with the model’s own interleaved positional settings. A factor of four targets the one-million-token workflow; for a workload around 524,288 tokens, the maker suggests a factor of two. Static scaling can affect shorter-text quality, so Qwen advises enabling it when the longer context is needed. Do not copy a different Qwen-VL edition’s extrapolation factor without its own recipe.
Keep long input organized by document, page or code identifier and request a checkable result. Verify calculations, small visual labels and generated patches against the source. Long context, retained reasoning and larger video budgets help carry more evidence through a task; they do not by themselves establish complete recall, correct reasoning or a completed external action.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3.8-27b | 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 $0.00085 / 1K tokens Output · Output tokens $0.0051 / 1K tokens Input · Cache read $0.00017 / 1K tokens Input · Cache write $0.0010625 / 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 $0.00085 / 1K tokens Output · Output tokens $0.0051 / 1K tokens Input · Cache read $0.00017 / 1K tokens Input · Cache write $0.0010625 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3.8-27b",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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/qwen3.8-27b', '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/qwen3.8-27b', '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/qwen3.8-27b
Use the method paths shown above.
Technical example in English
Analyze a circuit described in text\n\nA written circuit description gives two 6-ohm resistors connected in parallel across an ideal 12-volt source. Assume ideal wires and constant resistance. Derive the equivalent resistance, total current and current through each branch. Show the formulas and verify that branch currents add to the total. No circuit image or execution tool is provided.A short derivation with units, three requested quantities and a current-sum check.Equivalent resistance is 3 ohms and the two equal branch currents sum to the derived total current.