DeepSeek-R1-Distill-Qwen-32B is DeepSeek's 32B-size dense reasoning checkpoint built from Qwen2.5-32B. DeepSeek fine-tunes the base with reasoning examples generated by DeepSeek-R1. The result is a distinct text model for problems that require several connected steps, such as a mathematical derivation, an algorithmic question or reasoning through constraints.
How this distillation differs from the full R1 model
The maker's release separates DeepSeek-R1-Zero, DeepSeek-R1 and distilled models based on Qwen or Llama. The full R1 line develops reasoning through its own training pipeline; the distillation stage transfers patterns represented in R1-generated examples into separate dense checkpoints. This entry is the Qwen-based 32B distill edition, not the full R1 model, R1-Zero or the unmodified Qwen2.5-32B base.
The released Qwen distill models are fine-tuned with 800,000 curated samples. DeepSeek also changes their configurations and tokenizers and instructs users to keep those settings. The training origin explains the reasoning focus, while the named checkpoint and its own configuration determine the actual release. It does not mean that the larger teacher model is automatically consulted to answer each new question.
The 32B checkpoint's configuration uses a causal language-model architecture with 64 layers, 40 query attention heads and 8 key/value heads. It sets maximum positions to 131,072. This configuration figure concerns the released model; an example serving command or another member of the R1 family should not be treated as the same context setting. The model receives text and generates text rather than accepting an image or producing a visual or audio artifact.
Preparing a mathematical or coding task
For a mathematical problem, give the complete condition, define the variables and state the domain or assumptions. Identify whether the final result should be an exact expression, a numerical approximation or a proof of a claim. A useful request can ask for a derivation followed by the final expression and a check against the original conditions. The maker recommends a worked mathematical response with the final answer enclosed in a LaTeX box.
For an algorithm problem, provide the input range, required output and constraints on the solution. Ask for the approach, the implementation and an explanation of why it preserves the required property. If the problem concerns existing code, supply the relevant function, the input that produces the unexpected behavior and the expected result. The response can compare the actual path with the intended one and propose a correction; execution and verification of the program are separate application steps.
For a constraint problem, list the permitted choices and relationships explicitly. For example, a scheduling request can give the participants, availability and rules that make two assignments incompatible. Ask for a proposed assignment with the rule supporting each decision, and require any unresolved condition to remain visible. This makes the reasoning task inspectable instead of asking the model to choose without the underlying constraints.
Maker-specific usage recommendations
DeepSeek recommends putting all instructions in the user prompt and avoiding a separate system prompt for this series. This differs from a general chat template that relies on a system message to define the task. Place the requested answer format, the source material and the actual question together with clear boundaries.
The recommended temperature range is 0.5 to 0.7, with 0.6 preferred. The maker relates that recommendation to avoiding repetition or incoherent output. It also notes that some queries can produce an empty reasoning block and recommends beginning the output with the opening reasoning marker. These are recommendations for the released reasoning workflow, not a guarantee that one setting establishes correctness.
A practical answer contract can request an explanation of the method, the important justified steps and a separate final artifact. In a mathematical task, use the boxed expression for the final result. In a coding task, separate the code from its assumptions and example inputs. For a logical task, keep each conclusion connected to the rule or supplied fact that supports it. This allows the reasoning text and the result needed by the application to serve distinct purposes.
Context, output and integration boundaries
The released configuration's 131,072 maximum positions should be considered together with the full prompt and generated reasoning. The maker uses a 32,768-token generation setting in evaluation, but that experimental setting is not a published, separate output-capacity specification for this checkpoint. Do not derive an answer ceiling from a benchmark setup or copy the 8,192-token output figure of the original Qwen Instruct edition into the modified distill release.
Keep the inputs focused on the actual mathematical, programming or logical question, and state the desired final form. An external calculator, repository, document parser or search function needs to be supplied by the surrounding application; it is not included merely by selecting a text reasoning checkpoint. The useful result is a written solution or proposed implementation whose assumptions, method and final answer correspond to the supplied problem.
", "aboutText": "DeepSeek-R1-Distill-Qwen-32B is DeepSeek's 32B-size dense reasoning checkpoint built from Qwen2.5-32B. DeepSeek fine-tunes the base with reasoning examples generated by DeepSeek-R1. The result is a distinct text model for problems that require several connected steps, such as a mathematical derivation, an algorithmic question or reasoning through constraints.\n\nHow this distillation differs from the full R1 model\n\nThe maker's release separates DeepSeek-R1-Zero, DeepSeek-R1 and distilled models based on Qwen or Llama. The full R1 line develops reasoning through its own training pipeline; the distillation stage transfers patterns represented in R1-generated examples into separate dense checkpoints. This entry is the Qwen-based 32B distill edition, not the full R1 model, R1-Zero or the unmodified Qwen2.5-32B base.\n\nThe released Qwen distill models are fine-tuned with 800,000 curated samples. DeepSeek also changes their configurations and tokenizers and instructs users to keep those settings. The training origin explains the reasoning focus, while the named checkpoint and its own configuration determine the actual release. It does not mean that the larger teacher model is automatically consulted to answer each new question.\n\nThe 32B checkpoint's configuration uses a causal language-model architecture with 64 layers, 40 query attention heads and 8 key/value heads. It sets maximum positions to 131,072. This configuration figure concerns the released model; an example serving command or another member of the R1 family should not be treated as the same context setting. The model receives text and generates text rather than accepting an image or producing a visual or audio artifact.\n\nPreparing a mathematical or coding task\n\nFor a mathematical problem, give the complete condition, define the variables and state the domain or assumptions. Identify whether the final result should be an exact expression, a numerical approximation or a proof of a claim. A useful request can ask for a derivation followed by the final expression and a check against the original conditions. The maker recommends a worked mathematical response with the final answer enclosed in a LaTeX box.\n\nFor an algorithm problem, provide the input range, required output and constraints on the solution. Ask for the approach, the implementation and an explanation of why it preserves the required property. If the problem concerns existing code, supply the relevant function, the input that produces the unexpected behavior and the expected result. The response can compare the actual path with the intended one and propose a correction; execution and verification of the program are separate application steps.\n\nFor a constraint problem, list the permitted choices and relationships explicitly. For example, a scheduling request can give the participants, availability and rules that make two assignments incompatible. Ask for a proposed assignment with the rule supporting each decision, and require any unresolved condition to remain visible. This makes the reasoning task inspectable instead of asking the model to choose without the underlying constraints.\n\nMaker-specific usage recommendations\n\nDeepSeek recommends putting all instructions in the user prompt and avoiding a separate system prompt for this series. This differs from a general chat template that relies on a system message to define the task. Place the requested answer format, the source material and the actual question together with clear boundaries.\n\nThe recommended temperature range is 0.5 to 0.7, with 0.6 preferred. The maker relates that recommendation to avoiding repetition or incoherent output. It also notes that some queries can produce an empty reasoning block and recommends beginning the output with the opening reasoning marker. These are recommendations for the released reasoning workflow, not a guarantee that one setting establishes correctness.\n\nA practical answer contract can request an explanation of the method, the important justified steps and a separate final artifact. In a mathematical task, use the boxed expression for the final result. In a coding task, separate the code from its assumptions and example inputs. For a logical task, keep each conclusion connected to the rule or supplied fact that supports it. This allows the reasoning text and the result needed by the application to serve distinct purposes.\n\nContext, output and integration boundaries\n\nThe released configuration's 131,072 maximum positions should be considered together with the full prompt and generated reasoning. The maker uses a 32,768-token generation setting in evaluation, but that experimental setting is not a published, separate output-capacity specification for this checkpoint. Do not derive an answer ceiling from a benchmark setup or copy the 8,192-token output figure of the original Qwen Instruct edition into the modified distill release.\n\nKeep the inputs focused on the actual mathematical, programming or logical question, and state the desired final form. An external calculator, repository, document parser or search function needs to be supplied by the surrounding application; it is not included merely by selecting a text reasoning checkpoint. The useful result is a written solution or proposed implementation whose assumptions, method and final answer correspond to the supplied problem.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Solve a bounded scheduling problem\n\nSchedule three one-hour interviews in slots 09:00, 10:00 and 11:00. Ada cannot attend 09:00; Ben cannot attend 11:00; Chen must be before Ada. Each person gets one distinct slot. Give a valid schedule, explain how each constraint is satisfied and state whether another schedule is possible.A person-to-slot table, a constraint check and a justified uniqueness or non-uniqueness conclusion.Check all three constraints and distinct slots directly against the proposed schedule.deepseek-ai/DeepSeek-R1-Distill-Qwen-32BDeepSeek-R1-Distill-Qwen-32B is a dense text reasoning model fine-tuned from Qwen2.5-32B with DeepSeek-R1 examples. It is intended for multi-step mathematical, coding and logical tasks, with maker-specific prompting and a 131,072-position release configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
DeepSeek-R1-Distill-Qwen-32B is DeepSeek's 32B-size dense reasoning checkpoint built from Qwen2.5-32B. DeepSeek fine-tunes the base with reasoning examples generated by DeepSeek-R1. The result is a distinct text model for problems that require several connected steps, such as a mathematical derivation, an algorithmic question or reasoning through constraints.
How this distillation differs from the full R1 model
The maker's release separates DeepSeek-R1-Zero, DeepSeek-R1 and distilled models based on Qwen or Llama. The full R1 line develops reasoning through its own training pipeline; the distillation stage transfers patterns represented in R1-generated examples into separate dense checkpoints. This entry is the Qwen-based 32B distill edition, not the full R1 model, R1-Zero or the unmodified Qwen2.5-32B base.
The released Qwen distill models are fine-tuned with 800,000 curated samples. DeepSeek also changes their configurations and tokenizers and instructs users to keep those settings. The training origin explains the reasoning focus, while the named checkpoint and its own configuration determine the actual release. It does not mean that the larger teacher model is automatically consulted to answer each new question.
The 32B checkpoint's configuration uses a causal language-model architecture with 64 layers, 40 query attention heads and 8 key/value heads. It sets maximum positions to 131,072. This configuration figure concerns the released model; an example serving command or another member of the R1 family should not be treated as the same context setting. The model receives text and generates text rather than accepting an image or producing a visual or audio artifact.
Preparing a mathematical or coding task
For a mathematical problem, give the complete condition, define the variables and state the domain or assumptions. Identify whether the final result should be an exact expression, a numerical approximation or a proof of a claim. A useful request can ask for a derivation followed by the final expression and a check against the original conditions. The maker recommends a worked mathematical response with the final answer enclosed in a LaTeX box.
For an algorithm problem, provide the input range, required output and constraints on the solution. Ask for the approach, the implementation and an explanation of why it preserves the required property. If the problem concerns existing code, supply the relevant function, the input that produces the unexpected behavior and the expected result. The response can compare the actual path with the intended one and propose a correction; execution and verification of the program are separate application steps.
For a constraint problem, list the permitted choices and relationships explicitly. For example, a scheduling request can give the participants, availability and rules that make two assignments incompatible. Ask for a proposed assignment with the rule supporting each decision, and require any unresolved condition to remain visible. This makes the reasoning task inspectable instead of asking the model to choose without the underlying constraints.
Maker-specific usage recommendations
DeepSeek recommends putting all instructions in the user prompt and avoiding a separate system prompt for this series. This differs from a general chat template that relies on a system message to define the task. Place the requested answer format, the source material and the actual question together with clear boundaries.
The recommended temperature range is 0.5 to 0.7, with 0.6 preferred. The maker relates that recommendation to avoiding repetition or incoherent output. It also notes that some queries can produce an empty reasoning block and recommends beginning the output with the opening reasoning marker. These are recommendations for the released reasoning workflow, not a guarantee that one setting establishes correctness.
A practical answer contract can request an explanation of the method, the important justified steps and a separate final artifact. In a mathematical task, use the boxed expression for the final result. In a coding task, separate the code from its assumptions and example inputs. For a logical task, keep each conclusion connected to the rule or supplied fact that supports it. This allows the reasoning text and the result needed by the application to serve distinct purposes.
Context, output and integration boundaries
The released configuration's 131,072 maximum positions should be considered together with the full prompt and generated reasoning. The maker uses a 32,768-token generation setting in evaluation, but that experimental setting is not a published, separate output-capacity specification for this checkpoint. Do not derive an answer ceiling from a benchmark setup or copy the 8,192-token output figure of the original Qwen Instruct edition into the modified distill release.
Keep the inputs focused on the actual mathematical, programming or logical question, and state the desired final form. An external calculator, repository, document parser or search function needs to be supplied by the surrounding application; it is not included merely by selecting a text reasoning checkpoint. The useful result is a written solution or proposed implementation whose assumptions, method and final answer correspond to the supplied problem.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits deepseek-ai/DeepSeek-R1-Distill-Qwen-32B | 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 $0.00036 / 1K tokens Output · Output tokens $0.00036 / 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": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B", "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='deepseek-ai/DeepSeek-R1-Distill-Qwen-32B',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='deepseek-ai/DeepSeek-R1-Distill-Qwen-32B',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 deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
Use the method paths shown above.
deepseek-ai/DeepSeek-R1-Distill-Qwen-32BDeepSeek-R1-Distill-Qwen-32B is a dense text reasoning model fine-tuned from Qwen2.5-32B with DeepSeek-R1 examples. It is intended for multi-step mathematical, coding and logical tasks, with maker-specific prompting and a 131,072-position release configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
DeepSeek-R1-Distill-Qwen-32B is DeepSeek's 32B-size dense reasoning checkpoint built from Qwen2.5-32B. DeepSeek fine-tunes the base with reasoning examples generated by DeepSeek-R1. The result is a distinct text model for problems that require several connected steps, such as a mathematical derivation, an algorithmic question or reasoning through constraints.
How this distillation differs from the full R1 model
The maker's release separates DeepSeek-R1-Zero, DeepSeek-R1 and distilled models based on Qwen or Llama. The full R1 line develops reasoning through its own training pipeline; the distillation stage transfers patterns represented in R1-generated examples into separate dense checkpoints. This entry is the Qwen-based 32B distill edition, not the full R1 model, R1-Zero or the unmodified Qwen2.5-32B base.
The released Qwen distill models are fine-tuned with 800,000 curated samples. DeepSeek also changes their configurations and tokenizers and instructs users to keep those settings. The training origin explains the reasoning focus, while the named checkpoint and its own configuration determine the actual release. It does not mean that the larger teacher model is automatically consulted to answer each new question.
The 32B checkpoint's configuration uses a causal language-model architecture with 64 layers, 40 query attention heads and 8 key/value heads. It sets maximum positions to 131,072. This configuration figure concerns the released model; an example serving command or another member of the R1 family should not be treated as the same context setting. The model receives text and generates text rather than accepting an image or producing a visual or audio artifact.
Preparing a mathematical or coding task
For a mathematical problem, give the complete condition, define the variables and state the domain or assumptions. Identify whether the final result should be an exact expression, a numerical approximation or a proof of a claim. A useful request can ask for a derivation followed by the final expression and a check against the original conditions. The maker recommends a worked mathematical response with the final answer enclosed in a LaTeX box.
For an algorithm problem, provide the input range, required output and constraints on the solution. Ask for the approach, the implementation and an explanation of why it preserves the required property. If the problem concerns existing code, supply the relevant function, the input that produces the unexpected behavior and the expected result. The response can compare the actual path with the intended one and propose a correction; execution and verification of the program are separate application steps.
For a constraint problem, list the permitted choices and relationships explicitly. For example, a scheduling request can give the participants, availability and rules that make two assignments incompatible. Ask for a proposed assignment with the rule supporting each decision, and require any unresolved condition to remain visible. This makes the reasoning task inspectable instead of asking the model to choose without the underlying constraints.
Maker-specific usage recommendations
DeepSeek recommends putting all instructions in the user prompt and avoiding a separate system prompt for this series. This differs from a general chat template that relies on a system message to define the task. Place the requested answer format, the source material and the actual question together with clear boundaries.
The recommended temperature range is 0.5 to 0.7, with 0.6 preferred. The maker relates that recommendation to avoiding repetition or incoherent output. It also notes that some queries can produce an empty reasoning block and recommends beginning the output with the opening reasoning marker. These are recommendations for the released reasoning workflow, not a guarantee that one setting establishes correctness.
A practical answer contract can request an explanation of the method, the important justified steps and a separate final artifact. In a mathematical task, use the boxed expression for the final result. In a coding task, separate the code from its assumptions and example inputs. For a logical task, keep each conclusion connected to the rule or supplied fact that supports it. This allows the reasoning text and the result needed by the application to serve distinct purposes.
Context, output and integration boundaries
The released configuration's 131,072 maximum positions should be considered together with the full prompt and generated reasoning. The maker uses a 32,768-token generation setting in evaluation, but that experimental setting is not a published, separate output-capacity specification for this checkpoint. Do not derive an answer ceiling from a benchmark setup or copy the 8,192-token output figure of the original Qwen Instruct edition into the modified distill release.
Keep the inputs focused on the actual mathematical, programming or logical question, and state the desired final form. An external calculator, repository, document parser or search function needs to be supplied by the surrounding application; it is not included merely by selecting a text reasoning checkpoint. The useful result is a written solution or proposed implementation whose assumptions, method and final answer correspond to the supplied problem.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits deepseek-ai/DeepSeek-R1-Distill-Qwen-32B | 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 $0.00036 / 1K tokens Output · Output tokens $0.00036 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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': 'deepseek-ai/DeepSeek-R1-Distill-Qwen-32B', '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': 'deepseek-ai/DeepSeek-R1-Distill-Qwen-32B', '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 deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
Use the method paths shown above.
Technical example in English
Solve a bounded scheduling problem\n\nSchedule three one-hour interviews in slots 09:00, 10:00 and 11:00. Ada cannot attend 09:00; Ben cannot attend 11:00; Chen must be before Ada. Each person gets one distinct slot. Give a valid schedule, explain how each constraint is satisfied and state whether another schedule is possible.A person-to-slot table, a constraint check and a justified uniqueness or non-uniqueness conclusion.Check all three constraints and distinct slots directly against the proposed schedule.