Qwen3-Next highlights: complex text reasoning
Qwen3-Next-80B-A3B-Thinking is the thinking-only text edition of Qwen3-Next. The maker recommends it for highly complex reasoning tasks and notes that its thinking can be longer than in earlier models. It accepts text and returns text, including an explained answer or code proposal. Keep it distinct from Qwen3-Next-80B-A3B-Instruct and Qwen3-Coder-Next, whose non-thinking contracts are different.
The family introduces hybrid attention combining Gated DeltaNet and gated attention, high-sparsity experts, stability techniques such as zero-centered and weight-decayed layer normalization, and multi-token prediction. Qwen also describes reinforcement-learning stabilization through GSPO. These are training and architecture features; efficiency gains depend on implementation and should not become a universal response-time claim.
Model overview
The model has 80 billion parameters in total and activates three billion. It has 48 layers in a three-DeltaNet-to-one-gated-attention expert pattern. The maker lists 512 experts, ten activated experts and one shared expert, with pretraining on 15 trillion tokens followed by post-training. Activated parameters are not the complete checkpoint size.
The native context is 262,144 tokens, with documented extension to 1,010,000. Source text, historical messages, tool observations, reasoning and final output share the configured sequence allowance. For complex thinking, the maker advises a configured context greater than 131,072 where possible, even if a smaller window would reduce memory pressure.
Using the thinking-only interface
The default chat template already includes the opening thinking marker. A returned sequence can therefore contain the closing marker without repeating the opening one. This is normal formatting for the maker’s recipe, not evidence that the reasoning is missing. The model supports only thinking mode; a concise final-answer request does not convert it into a non-thinking edition.
For a mathematical or logical problem, give the full statement, definitions, units and allowed assumptions. Ask for a clearly delimited final answer with the supporting steps needed to check it. For code reasoning, provide the relevant function, input and intended behavior, then request an explained proposal. Compare the supporting calculations or proposed patch with the supplied conditions and actual results.
Implementation and tools
The maker’s examples support development and tool workflows and recommend Qwen-Agent for tool templates and parsing, with configured or MCP-connected tools. Define the permitted tools, task goal and observable completion condition. The application executes external actions and returns actual observations; the model can then use that feedback in its next answer.
Multi-token prediction is part of training, but its inference use depends on implementation and is not generally available in every lightweight workflow. Preserve that distinction when assessing a deployment. Tool or architecture capability alone does not establish access to files, current sources or an execution environment through another application.
Ultra-long text
The documented extension uses YaRN. The maker validates approximately one-million-token context and provides a factor-of-four recipe based on the native 262,144 positions; for 524,288-token workloads it suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer input.
Organize a large input with document boundaries and code or section identifiers. State the question and expected artifact so the model can connect relevant evidence rather than treat the full window as a substitute for a clear task. Reserve space for its potentially long reasoning and the final result.
Generation guidance
Best Practices recommend temperature 0.6, top_p 0.95, top_k 20 and min_p 0. Where supported, a presence penalty between zero and two can reduce repeated text, but higher values may mix languages or slightly reduce quality. Qwen recommends a 32,768-token generation allowance for most questions and 81,920 for unusually complex math or programming tasks. These are recommendations rather than independent hard output ceilings.
Specify the final-answer format explicitly, such as a delimited mathematical result or an answer field. The maker recommends keeping final outputs without thinking content in historical assistant messages, following this checkpoint’s template. Later models’ historical reasoning-preservation options should not be imported automatically.
Substantive limits
This is a text-only model. Images, audio or video require an appropriate separate representation before they can serve as evidence in this workflow. Long thinking and context are finite resources, and detailed reasoning does not independently prove a correct result. Check assumptions, calculations and generated code against the supplied source. The useful outcome is a reviewable answer or artifact whose evidence remains visible and whose thinking-only identity is preserved.
", "aboutText": "Qwen3-Next highlights: complex text reasoning\n\nQwen3-Next-80B-A3B-Thinking is the thinking-only text edition of Qwen3-Next. The maker recommends it for highly complex reasoning tasks and notes that its thinking can be longer than in earlier models. It accepts text and returns text, including an explained answer or code proposal. Keep it distinct from Qwen3-Next-80B-A3B-Instruct and Qwen3-Coder-Next, whose non-thinking contracts are different.\n\nThe family introduces hybrid attention combining Gated DeltaNet and gated attention, high-sparsity experts, stability techniques such as zero-centered and weight-decayed layer normalization, and multi-token prediction. Qwen also describes reinforcement-learning stabilization through GSPO. These are training and architecture features; efficiency gains depend on implementation and should not become a universal response-time claim.\n\nModel overview\n\nThe model has 80 billion parameters in total and activates three billion. It has 48 layers in a three-DeltaNet-to-one-gated-attention expert pattern. The maker lists 512 experts, ten activated experts and one shared expert, with pretraining on 15 trillion tokens followed by post-training. Activated parameters are not the complete checkpoint size.\n\nThe native context is 262,144 tokens, with documented extension to 1,010,000. Source text, historical messages, tool observations, reasoning and final output share the configured sequence allowance. For complex thinking, the maker advises a configured context greater than 131,072 where possible, even if a smaller window would reduce memory pressure.\n\nUsing the thinking-only interface\n\nThe default chat template already includes the opening thinking marker. A returned sequence can therefore contain the closing marker without repeating the opening one. This is normal formatting for the maker’s recipe, not evidence that the reasoning is missing. The model supports only thinking mode; a concise final-answer request does not convert it into a non-thinking edition.\n\nFor a mathematical or logical problem, give the full statement, definitions, units and allowed assumptions. Ask for a clearly delimited final answer with the supporting steps needed to check it. For code reasoning, provide the relevant function, input and intended behavior, then request an explained proposal. Compare the supporting calculations or proposed patch with the supplied conditions and actual results.\n\nImplementation and tools\n\nThe maker’s examples support development and tool workflows and recommend Qwen-Agent for tool templates and parsing, with configured or MCP-connected tools. Define the permitted tools, task goal and observable completion condition. The application executes external actions and returns actual observations; the model can then use that feedback in its next answer.\n\nMulti-token prediction is part of training, but its inference use depends on implementation and is not generally available in every lightweight workflow. Preserve that distinction when assessing a deployment. Tool or architecture capability alone does not establish access to files, current sources or an execution environment through another application.\n\nUltra-long text\n\nThe documented extension uses YaRN. The maker validates approximately one-million-token context and provides a factor-of-four recipe based on the native 262,144 positions; for 524,288-token workloads it suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer input.\n\nOrganize a large input with document boundaries and code or section identifiers. State the question and expected artifact so the model can connect relevant evidence rather than treat the full window as a substitute for a clear task. Reserve space for its potentially long reasoning and the final result.\n\nGeneration guidance\n\nBest Practices recommend temperature 0.6, top_p 0.95, top_k 20 and min_p 0. Where supported, a presence penalty between zero and two can reduce repeated text, but higher values may mix languages or slightly reduce quality. Qwen recommends a 32,768-token generation allowance for most questions and 81,920 for unusually complex math or programming tasks. These are recommendations rather than independent hard output ceilings.\n\nSpecify the final-answer format explicitly, such as a delimited mathematical result or an answer field. The maker recommends keeping final outputs without thinking content in historical assistant messages, following this checkpoint’s template. Later models’ historical reasoning-preservation options should not be imported automatically.\n\nSubstantive limits\n\nThis is a text-only model. Images, audio or video require an appropriate separate representation before they can serve as evidence in this workflow. Long thinking and context are finite resources, and detailed reasoning does not independently prove a correct result. Check assumptions, calculations and generated code against the supplied source. The useful outcome is a reviewable answer or artifact whose evidence remains visible and whose thinking-only identity is preserved.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Identify what a logic puzzle actually determines\n\nExactly one of A, B and C took a key, and exactly one statement is true. A says, 'B took it.' B says, 'C took it.' C says, 'I did not take it.' Determine every possible culprit consistent with these premises. Show a truth table and state whether the culprit is uniquely determined.A three-case truth table and the complete set of consistent culprits with a uniqueness conclusion.The table checks exactly one true statement in each candidate case and preserves multiple solutions if the premises allow them.qwen/qwen3-next-80b-a3b-thinkingQwen3-Next-80B-A3B-Thinking is Qwen’s text reasoning model with 80 billion total and 3 billion activated parameters. It uses thinking only, hybrid attention and sparse experts for complex tasks, with a native 262,144-token context and conditional extension to 1,010,000.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3-Next highlights: complex text reasoning
Qwen3-Next-80B-A3B-Thinking is the thinking-only text edition of Qwen3-Next. The maker recommends it for highly complex reasoning tasks and notes that its thinking can be longer than in earlier models. It accepts text and returns text, including an explained answer or code proposal. Keep it distinct from Qwen3-Next-80B-A3B-Instruct and Qwen3-Coder-Next, whose non-thinking contracts are different.
The family introduces hybrid attention combining Gated DeltaNet and gated attention, high-sparsity experts, stability techniques such as zero-centered and weight-decayed layer normalization, and multi-token prediction. Qwen also describes reinforcement-learning stabilization through GSPO. These are training and architecture features; efficiency gains depend on implementation and should not become a universal response-time claim.
Model overview
The model has 80 billion parameters in total and activates three billion. It has 48 layers in a three-DeltaNet-to-one-gated-attention expert pattern. The maker lists 512 experts, ten activated experts and one shared expert, with pretraining on 15 trillion tokens followed by post-training. Activated parameters are not the complete checkpoint size.
The native context is 262,144 tokens, with documented extension to 1,010,000. Source text, historical messages, tool observations, reasoning and final output share the configured sequence allowance. For complex thinking, the maker advises a configured context greater than 131,072 where possible, even if a smaller window would reduce memory pressure.
Using the thinking-only interface
The default chat template already includes the opening thinking marker. A returned sequence can therefore contain the closing marker without repeating the opening one. This is normal formatting for the maker’s recipe, not evidence that the reasoning is missing. The model supports only thinking mode; a concise final-answer request does not convert it into a non-thinking edition.
For a mathematical or logical problem, give the full statement, definitions, units and allowed assumptions. Ask for a clearly delimited final answer with the supporting steps needed to check it. For code reasoning, provide the relevant function, input and intended behavior, then request an explained proposal. Compare the supporting calculations or proposed patch with the supplied conditions and actual results.
Implementation and tools
The maker’s examples support development and tool workflows and recommend Qwen-Agent for tool templates and parsing, with configured or MCP-connected tools. Define the permitted tools, task goal and observable completion condition. The application executes external actions and returns actual observations; the model can then use that feedback in its next answer.
Multi-token prediction is part of training, but its inference use depends on implementation and is not generally available in every lightweight workflow. Preserve that distinction when assessing a deployment. Tool or architecture capability alone does not establish access to files, current sources or an execution environment through another application.
Ultra-long text
The documented extension uses YaRN. The maker validates approximately one-million-token context and provides a factor-of-four recipe based on the native 262,144 positions; for 524,288-token workloads it suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer input.
Organize a large input with document boundaries and code or section identifiers. State the question and expected artifact so the model can connect relevant evidence rather than treat the full window as a substitute for a clear task. Reserve space for its potentially long reasoning and the final result.
Generation guidance
Best Practices recommend temperature 0.6, top_p 0.95, top_k 20 and min_p 0. Where supported, a presence penalty between zero and two can reduce repeated text, but higher values may mix languages or slightly reduce quality. Qwen recommends a 32,768-token generation allowance for most questions and 81,920 for unusually complex math or programming tasks. These are recommendations rather than independent hard output ceilings.
Specify the final-answer format explicitly, such as a delimited mathematical result or an answer field. The maker recommends keeping final outputs without thinking content in historical assistant messages, following this checkpoint’s template. Later models’ historical reasoning-preservation options should not be imported automatically.
Substantive limits
This is a text-only model. Images, audio or video require an appropriate separate representation before they can serve as evidence in this workflow. Long thinking and context are finite resources, and detailed reasoning does not independently prove a correct result. Check assumptions, calculations and generated code against the supplied source. The useful outcome is a reviewable answer or artifact whose evidence remains visible and whose thinking-only identity is preserved.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-next-80b-a3b-thinking | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits Minimum items: 1 | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits Examples: system, user, assistant, tool, function Conditions: These roles are explicitly represented by Source compatibility types; the exact model role set remains model-specific. Model-specific | Default— |
Fieldmessages[].content | Typestring|content_part[]|null | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_calls | Typetool_call[] | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_call_id | Typestring | RequirementOptional | Values & limits Applies when: role=tool; use actual returned ID | Default— |
Fieldmessages[].name | Typestring | RequirementOptional | Values & limits | Default— |
Fieldmessages[].reasoning_content | Typestring | RequirementOptional | Values & limits Applies when: Round-trip actual returned reasoning where this model uses it | Default— |
Fieldmessages[].reasoning_details | Typeobject|array | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmax_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldmax_completion_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Fieldstream_options.include_usage | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtemperature | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtop_p | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldstop | Typestring|string[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools | Typefunction_tool[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools[].type | Typestring | RequirementOptional | Values & limits function Model-specific | Default— |
Fieldtools[].function.name | Typestring | RequirementRequired | Values & limits Applies when: function tool supplied | Default— |
Fieldtools[].function.description | Typestring | RequirementOptional | Values & limits | Default— |
Fieldtools[].function.parameters | TypeJSON_schema | RequirementOptional | Values & limits | Default— |
Fieldtool_choice | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldparallel_tool_calls | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldresponse_format | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning_effort | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldservice_tier | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldinstructions | Typestring | RequirementOptional | Values & limits Compatibility field | Default— |
Fieldfunctions | Typefunction[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldfunction_call | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldprompt_cache_options | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmodalities | Typestring[] | RequirementOptional | Values & limits text | Default— |
Fieldprovider | Typeobject | RequirementOptional | Values & limits Conditions: provider.options is rejected on this Ultra text route | Default— |
Fieldcache_control.ttl | Typestring | RequirementOptional | Values & limits Conditions: 1h is rejected in actual prompt-cache locations | Default— |
Text · UTF8 text
messages[].content
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages |
Parametermaterial_types | Contracttext |
Parametermessages | Contractmessage array with role/content shape |
Parameterfile_material | Contractactual UTF-8 document inserted as literal text; no file ID inferred |
Parameteroptional_controls | Contractstream, temperature, top_p, stop, max_tokens, max_completion_tokens, reasoning, response_format, tools, tool_choice |
Parametercontrol_scope | Contractmodel-specific; user function tools differ from hosted tools |
Parameteroutput_modalities | Contracttext |
Parameterhosted_operations | Contractfalse |
Parametermultimodal_input_binding | Contractnull |
Lifecycle: synchronous_or_streamed
Send exact model ID with messages.
Context length is not an output cap.
A text file may supply literal text; image/audio/file-object inputs need exact binding.
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0003 / 1K tokens Output · Output tokens $0.0024 / 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 $0.0003 / 1K tokens Output · Output tokens $0.0024 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "qwen/qwen3-next-80b-a3b-thinking", "messages": [{"role": "user", "content": "Explain the difference between an observation and a hypothesis in three sentences."}]}')\n\nPath("request.json").write_text(json.dumps(p,ensure_ascii=False))\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/chat/completions \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\nr=json.loads(Path("result.json").read_text())\nfor x in r.get("choices",[]):\n if x.get("message",{}).get("content"): print(x["message"]["content"])\nPY_RESULT\nfrom openai import OpenAI\nimport os\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nresult=client.chat.completions.create(model='qwen/qwen3-next-80b-a3b-thinking',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}])\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nsave_result(result.to_dict(mode="json"))\nfrom openai import OpenAI\nimport os\nfrom pathlib import Path\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nwith client.chat.completions.stream(model='qwen/qwen3-next-80b-a3b-thinking',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}]) as stream:\n with Path("stream-events.jsonl").open("w",encoding="utf-8") as log:\n for event in stream:\n payload=event.to_dict(mode="json")\n log.write(json.dumps(payload,ensure_ascii=False)+"\\n"); log.flush()\n print(json.dumps(payload,ensure_ascii=False))\n final=stream.get_final_completion()\n save_result(final.to_dict(mode="json"),"stream-result")\njson or SSE
Keep every complete returned response, file and content part.
Keep every choice and its content/tool calls/reasoning. In SSE assemble deltas by choice/tool index, retain the final finish reason and usage.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldchoices[].index | Typeinteger |
Fieldchoices[].message.content | Typestring|null |
Fieldchoices[].message.tool_calls | Typetool_call[] |
Fieldchoices[].message.reasoning_content | Typestring |
Fieldchoices[].message.reasoning_details | Typeobject|array |
Fieldchoices[].finish_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen3-next-80b-a3b-thinking
Use the method paths shown above.
qwen/qwen3-next-80b-a3b-thinkingQwen3-Next-80B-A3B-Thinking is Qwen’s text reasoning model with 80 billion total and 3 billion activated parameters. It uses thinking only, hybrid attention and sparse experts for complex tasks, with a native 262,144-token context and conditional extension to 1,010,000.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3-Next highlights: complex text reasoning
Qwen3-Next-80B-A3B-Thinking is the thinking-only text edition of Qwen3-Next. The maker recommends it for highly complex reasoning tasks and notes that its thinking can be longer than in earlier models. It accepts text and returns text, including an explained answer or code proposal. Keep it distinct from Qwen3-Next-80B-A3B-Instruct and Qwen3-Coder-Next, whose non-thinking contracts are different.
The family introduces hybrid attention combining Gated DeltaNet and gated attention, high-sparsity experts, stability techniques such as zero-centered and weight-decayed layer normalization, and multi-token prediction. Qwen also describes reinforcement-learning stabilization through GSPO. These are training and architecture features; efficiency gains depend on implementation and should not become a universal response-time claim.
Model overview
The model has 80 billion parameters in total and activates three billion. It has 48 layers in a three-DeltaNet-to-one-gated-attention expert pattern. The maker lists 512 experts, ten activated experts and one shared expert, with pretraining on 15 trillion tokens followed by post-training. Activated parameters are not the complete checkpoint size.
The native context is 262,144 tokens, with documented extension to 1,010,000. Source text, historical messages, tool observations, reasoning and final output share the configured sequence allowance. For complex thinking, the maker advises a configured context greater than 131,072 where possible, even if a smaller window would reduce memory pressure.
Using the thinking-only interface
The default chat template already includes the opening thinking marker. A returned sequence can therefore contain the closing marker without repeating the opening one. This is normal formatting for the maker’s recipe, not evidence that the reasoning is missing. The model supports only thinking mode; a concise final-answer request does not convert it into a non-thinking edition.
For a mathematical or logical problem, give the full statement, definitions, units and allowed assumptions. Ask for a clearly delimited final answer with the supporting steps needed to check it. For code reasoning, provide the relevant function, input and intended behavior, then request an explained proposal. Compare the supporting calculations or proposed patch with the supplied conditions and actual results.
Implementation and tools
The maker’s examples support development and tool workflows and recommend Qwen-Agent for tool templates and parsing, with configured or MCP-connected tools. Define the permitted tools, task goal and observable completion condition. The application executes external actions and returns actual observations; the model can then use that feedback in its next answer.
Multi-token prediction is part of training, but its inference use depends on implementation and is not generally available in every lightweight workflow. Preserve that distinction when assessing a deployment. Tool or architecture capability alone does not establish access to files, current sources or an execution environment through another application.
Ultra-long text
The documented extension uses YaRN. The maker validates approximately one-million-token context and provides a factor-of-four recipe based on the native 262,144 positions; for 524,288-token workloads it suggests a factor of two. Static scaling can affect shorter-text performance, so it should follow the actual need for longer input.
Organize a large input with document boundaries and code or section identifiers. State the question and expected artifact so the model can connect relevant evidence rather than treat the full window as a substitute for a clear task. Reserve space for its potentially long reasoning and the final result.
Generation guidance
Best Practices recommend temperature 0.6, top_p 0.95, top_k 20 and min_p 0. Where supported, a presence penalty between zero and two can reduce repeated text, but higher values may mix languages or slightly reduce quality. Qwen recommends a 32,768-token generation allowance for most questions and 81,920 for unusually complex math or programming tasks. These are recommendations rather than independent hard output ceilings.
Specify the final-answer format explicitly, such as a delimited mathematical result or an answer field. The maker recommends keeping final outputs without thinking content in historical assistant messages, following this checkpoint’s template. Later models’ historical reasoning-preservation options should not be imported automatically.
Substantive limits
This is a text-only model. Images, audio or video require an appropriate separate representation before they can serve as evidence in this workflow. Long thinking and context are finite resources, and detailed reasoning does not independently prove a correct result. Check assumptions, calculations and generated code against the supplied source. The useful outcome is a reviewable answer or artifact whose evidence remains visible and whose thinking-only identity is preserved.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-next-80b-a3b-thinking | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits user, assistant | Default— |
Fieldmessages[].content | Typestring | RequirementRequired | Values & limits | Default— |
Fieldmax_tokens | Typeinteger | RequirementRequired | Values & limits | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Text · UTF-8 text
messages[].content string
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages, max_tokens |
Parameteroptional_controls | Contractstream |
Parametermaterial_types | Contracttext |
Parametertext_file_material | ContractRead the actual UTF-8 document and put its text in the indicated input or message content string. |
Lifecycle: synchronous_or_streamed
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0003 / 1K tokens Output · Output tokens $0.0024 / 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 $0.0003 / 1K tokens Output · Output tokens $0.0024 / 1K tokens Input · Cache read — / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3-next-80b-a3b-thinking",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\n: "${AILOOT_TEXT_FILE:?Set your actual UTF-8 document}"\nexport AILOOT_TEXT_FILE\npython3 - <<'PY_REQUEST'\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen3-next-80b-a3b-thinking', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nPath('request.json').write_text(json.dumps(payload, ensure_ascii=False), encoding='utf-8')\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/messages \\\n -H "x-api-key: $AILOOT_API_KEY" -H 'anthropic-version: 2023-06-01' -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nPY_RESULT\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen3-next-80b-a3b-thinking', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nfrom urllib.request import Request, urlopen\nfrom urllib.error import HTTPError\n\napi_key = os.environ['AILOOT_API_KEY']\nif not api_key.strip():\n raise SystemExit('Set AILOOT_API_KEY before making a request')\nheaders = {'x-api-key': 'PLACEHOLDER_KEY', 'anthropic-version': '2023-06-01', 'Content-Type': 'application/json'}\nheaders['x-api-key'] = api_key\nrequest = Request('https://api.ailoot.ai/v1/messages', data=json.dumps(payload, ensure_ascii=False).encode('utf-8'), headers=headers, method='POST')\ntry:\n with urlopen(request, timeout=120) as response:\n raw = response.read()\n Path('response-headers.txt').write_text(str(response.headers), encoding='utf-8')\nexcept HTTPError as error:\n Path('result.json').write_bytes(error.read())\n Path('response-headers.txt').write_text(str(error.headers), encoding='utf-8')\n raise\nPath('result.json').write_bytes(raw)\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nJSON or SSE
Keep every complete returned response, file and content part.
Keep all content blocks, thinking signatures and actual tool-use IDs. Assemble all SSE block deltas; a first text block is not the complete output.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldid | Typestring |
Fieldmodel | Typestring |
Fieldcontent[] | Typetext|thinking|tool_use|other content block |
Fieldstop_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen3-next-80b-a3b-thinking
Use the method paths shown above.
Technical example in English
Identify what a logic puzzle actually determines\n\nExactly one of A, B and C took a key, and exactly one statement is true. A says, 'B took it.' B says, 'C took it.' C says, 'I did not take it.' Determine every possible culprit consistent with these premises. Show a truth table and state whether the culprit is uniquely determined.A three-case truth table and the complete set of consistent culprits with a uniqueness conclusion.The table checks exactly one true statement in each candidate case and preserves multiple solutions if the premises allow them.