Highlights: the fixed Thinking-2507 release
Qwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.
Model overview and mode
The mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.
The default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.
Preparing a difficult task
Give the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.
Because this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.
Agentic use
The maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.
Processing ultra-long text
Qwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.
That recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.
Generation guidance and limits
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.
Specify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.
", "aboutText": "Highlights: the fixed Thinking-2507 release\n\nQwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.\n\nModel overview and mode\n\nThe mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.\n\nThe default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.\n\nPreparing a difficult task\n\nGive the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.\n\nBecause this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.\n\nAgentic use\n\nThe maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.\n\nProcessing ultra-long text\n\nQwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.\n\nThat recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.\n\nGeneration guidance and limits\n\nQwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.\n\nSpecify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Solve a constrained staffing assignment\n\nAssign three reviewers, Ada, Ben and Cy, to morning, afternoon and evening shifts, one each. Ada cannot work evening. Ben must work later than Ada. Cy cannot work morning. List every valid assignment and explain how each constraint eliminates alternatives. Finish with whether the assignment is unique; do not assume additional availability.An enumeration of all valid assignments, a constraint check and a separate uniqueness conclusion.Every assignment uses all three shifts once and satisfies Ben later than Ada without adding availability rules.Qwen/Qwen3-30B-A3B-Thinking-2507Qwen3-30B-A3B-Thinking-2507 is Qwen’s thinking-only text model for demanding mathematics, science, coding and logical reasoning. This fixed 2507 edition provides a native 262,144-token context and a longer reasoning budget, with documented tool-calling and conditional ultra-long-context workflows.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Highlights: the fixed Thinking-2507 release
Qwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.
Model overview and mode
The mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.
The default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.
Preparing a difficult task
Give the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.
Because this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.
Agentic use
The maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.
Processing ultra-long text
Qwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.
That recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.
Generation guidance and limits
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.
Specify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits Qwen/Qwen3-30B-A3B-Thinking-2507 | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits Minimum items: 1 | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits Model-specific | Default— |
Fieldmessages[].content | Typestring | RequirementOptional | Values & limits Applies when: Text conversation | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
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 |
Parametercontrol_scope | ContractText conversation; stream is a boolean and defaults to false. |
Parameteroutput_modalities | Contracttext |
Parameterhosted_operations | Contractfalse |
Parametermultimodal_input_binding | Contractnull |
Method aliases: /chat/completions
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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.0008 / 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/Qwen3-30B-A3B-Thinking-2507","messages":[{"role":"user","content":"hello"}],"stream":false}')\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/Qwen3-30B-A3B-Thinking-2507", messages=[{"role":"user","content":"hello"}], stream=False)\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"))\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/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
/chat/completions
Qwen/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
Qwen/Qwen3-30B-A3B-Thinking-2507Qwen3-30B-A3B-Thinking-2507 is Qwen’s thinking-only text model for demanding mathematics, science, coding and logical reasoning. This fixed 2507 edition provides a native 262,144-token context and a longer reasoning budget, with documented tool-calling and conditional ultra-long-context workflows.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Highlights: the fixed Thinking-2507 release
Qwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.
Model overview and mode
The mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.
The default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.
Preparing a difficult task
Give the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.
Because this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.
Agentic use
The maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.
Processing ultra-long text
Qwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.
That recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.
Generation guidance and limits
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.
Specify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits Qwen/Qwen3-30B-A3B-Thinking-2507 | 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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.0008 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "Qwen/Qwen3-30B-A3B-Thinking-2507",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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/Qwen3-30B-A3B-Thinking-2507', '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/Qwen3-30B-A3B-Thinking-2507', '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/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
/chat/completions
Qwen/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
qwen3-30b-a3b-thinking-2507Qwen3-30B-A3B-Thinking-2507 is Qwen’s thinking-only text model for demanding mathematics, science, coding and logical reasoning. This fixed 2507 edition provides a native 262,144-token context and a longer reasoning budget, with documented tool-calling and conditional ultra-long-context workflows.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Highlights: the fixed Thinking-2507 release
Qwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.
Model overview and mode
The mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.
The default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.
Preparing a difficult task
Give the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.
Because this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.
Agentic use
The maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.
Processing ultra-long text
Qwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.
That recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.
Generation guidance and limits
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.
Specify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen3-30b-a3b-thinking-2507 | 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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.0008 / 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": "qwen3-30b-a3b-thinking-2507", "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='qwen3-30b-a3b-thinking-2507',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='qwen3-30b-a3b-thinking-2507',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/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
/chat/completions
Qwen/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
qwen3-30b-a3b-thinking-2507Qwen3-30B-A3B-Thinking-2507 is Qwen’s thinking-only text model for demanding mathematics, science, coding and logical reasoning. This fixed 2507 edition provides a native 262,144-token context and a longer reasoning budget, with documented tool-calling and conditional ultra-long-context workflows.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Highlights: the fixed Thinking-2507 release
Qwen3-30B-A3B-Thinking-2507 is a text-generation model focused on difficult reasoning. Qwen describes improvements in logic, mathematics, science, coding, instruction following, tool use and text generation, together with longer thinking sequences and stronger long-context understanding. Select this exact 2507 edition when the task benefits from sustained reasoning, such as examining competing explanations, solving a multi-stage mathematical problem or working through a complex code change.
Model overview and mode
The mixture-of-experts model has 30.5 billion total parameters and 3.3 billion activated parameters, with 128 experts and eight selected per token across 48 layers. Its native context is 262,144 tokens. It accepts text and returns text; tool results are provided through a connected application workflow. This checkpoint supports thinking only. It is distinct from the earlier hybrid Qwen3-30B-A3B, its FP8 edition and Qwen3-30B-A3B-Instruct-2507. The hybrid model’s non-thinking switch should not be assumed to apply here.
The default chat template already starts the reasoning sequence with an opening thinking marker, so an explicit enable_thinking=True setting is unnecessary. A returned sequence may contain the closing marker without repeating the opening marker. The maker’s example separates the reasoning from the final answer. Preserve that distinction when requesting a concrete deliverable: a proof conclusion, a patch or a decision table should remain easy to find after the longer explanation.
Preparing a difficult task
Give the model the complete problem statement, constraints and relevant evidence. For a code change, include the existing function, a failing input and expected behavior, then request a proposed patch with an explanation of edge cases. For a scientific question, separate observations from assumptions and ask which conclusion follows from each. For a long policy or specification, retain section identifiers and request a concise decision backed by the relevant clauses. These are practical uses of the documented reasoning domains, not claims that the model has executed a patch or validated an experiment.
Because this version can reason at length, reserve space for both reasoning and the final artifact. Qwen advises retaining a context above 131,072 tokens where possible when configuring the model, even if a smaller context would ease memory pressure. The native context figure includes the sequence budget; it is not a promise that all of it can be used for a document while leaving unlimited answer space.
Agentic use
The maker documents tool calling and recommends Qwen-Agent for tool templates and parsing, including integration with configured tools and code-interpreter workflows. Define the tools and the intended result explicitly. For example, an application could let the model inspect retrieved evidence, use a permitted calculator, then produce a report explaining the calculation. Retrieval and execution belong to the connected tools; the model’s tool-calling capability does not itself establish access to the internet, a code environment or business systems.
Processing ultra-long text
Qwen documents an optional workflow reaching approximately one million tokens by combining Dual Chunk Attention with MInference sparse attention. The first technique divides long sequences into chunks while preserving interactions across them; the second focuses computation on selected attention relationships. This is an additional configuration path, beyond the native 262,144-token context. The maker’s recipe requires its special long-context configuration and estimates roughly 240 GB of total GPU memory for weights, cached attention state and peak activations.
That recipe has concrete resource limits: cached sequence state, activation memory and total input length can each exhaust the configured capacity. A million-token workflow therefore depends on configuration, memory and supported attention implementation. For practical document review, start with the relevant material, mark document boundaries and request traceable evidence instead of treating the largest possible input as a substitute for a clear question.
Generation guidance and limits
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0. It suggests a 32,768-token output budget for most queries and 81,920 for unusually difficult mathematical or programming problems. These are generation-budget recommendations rather than a separately published hard output ceiling. A presence penalty between zero and two can reduce repetition, but higher values may mix languages or slightly lower quality.
Specify the final-answer format in the prompt, such as a delimited mathematical result or an answer field for a multiple-choice question. Keep only the final-answer portion in conversation history, as the maker recommends. Longer reasoning is useful when it produces a checkable result; inspect assumptions, calculations and proposed code before relying on the outcome.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen3-30b-a3b-thinking-2507 | 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.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.0008 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen3-30b-a3b-thinking-2507",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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': 'qwen3-30b-a3b-thinking-2507', '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': 'qwen3-30b-a3b-thinking-2507', '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/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
/chat/completions
Qwen/Qwen3-30B-A3B-Thinking-2507, qwen3-30b-a3b-thinking-2507
Technical example in English
Solve a constrained staffing assignment\n\nAssign three reviewers, Ada, Ben and Cy, to morning, afternoon and evening shifts, one each. Ada cannot work evening. Ben must work later than Ada. Cy cannot work morning. List every valid assignment and explain how each constraint eliminates alternatives. Finish with whether the assignment is unique; do not assume additional availability.An enumeration of all valid assignments, a constraint check and a separate uniqueness conclusion.Every assignment uses all three shifts once and satisfies Ben later than Ada without adding availability rules.