{ "aboutHtml": "

DeepSeek V3.2 是 DeepSeek 用於推理與代理工作的文字模型，著重高效率的長上下文處理。製造商將主要變更歸納為 DeepSeek Sparse Attention、擴展規模的強化學習框架，以及代理任務的訓練管線。它接受書面問題與上下文，並產生文字，包括答案、程式碼或工具呼叫內容。它與較早的 V3.2 Exp 版本及獨立的 Speciale 推理變體有所區別。

DeepSeek Sparse Attention 在潛在注意力架構中加入索引器與細粒度 token 選取。對長序列而言，索引器會選取主要注意力計算所用的先前 token 資訊。報告指出，這是相較於 V3.1 Terminus 的架構變更，且採用與 V3.2 Exp 相同的架構。公布的上下文長度為 128K token，可支援大量來源資料與持續進行的任務，但生成的推理與工具歷史也會占用可用上下文。

後訓練涵蓋數學、程式設計、邏輯推理、一般代理任務、代理式程式撰寫與搜尋中的思考及非思考回應。各領域專家提供資料，彙整至最終模型後，再進行混合強化學習。一般問題可能適合直接回答；困難問題則可透過思考工作流程，在呈現最終結果前發展解法。請將所需答案或成果與中間推理分開指定。

此版本的一項重大變更是結合工具進行思考。更新後的聊天格式讓推理能持續貫穿工具呼叫任務，而非將分析與操作視為無關的階段。模型選擇的操作由應用程式執行，再回傳結果供下一步使用。依製造商的工具上下文規則，只新增工具相關訊息時，會保留歷史推理；收到新的使用者訊息時，則會捨棄該推理，但保留工具呼叫與結果。正確區分對話角色會影響這項行為。

程式撰寫任務應提供原始碼片段或相關檔案、觀察到的問題、預期結果，以及修改必須保留的條件。任務說明範例為：「使用提供的程式碼與工具結果找出原因，提出最小修正，並回傳最終修補內容，以及支持該修正的驗證結果。」搜尋任務應說明問題、所需證據，以及何種結果才算完成回答。搜尋與執行工具由周邊應用程式提供，並非單靠文字即可取得。

製造商建議本機使用時將 temperature 設為 1.0、top-p 設為 0.95。V3.2 引入修訂的工具呼叫表示方式與專用訊息編碼器。其 developer 角色保留給製造商的搜尋代理範本，官方 API 不接受該角色。參考解析器要求格式正確的輸出，不會修復格式錯誤的結果；應用程式使用生成欄位或工具引數時，應先檢查結構再據此操作。

技術報告指出，過長的推理歷程與重複自我驗證是實際限制。即使原始來源資料能容納於上下文中，工具輸出與重複推理仍會累積，使代理任務達到 128K 邊界。請精簡記錄相關觀察，並明確指定最終成果。對知識準確性要求較高的工作，請附上權威資料或提供適當的搜尋工具，而非期待訓練知識能確定每項事實。

DeepSeek V3.2 Speciale 是獨立的高運算量深度推理變體；依模型卡所述，它不支援工具呼叫。不應以其功能描述標準 V3.2 的工具組態。同樣地，後來的 V4 世代與實驗性前代也各有不同身分與設定選項。比較模型與提示詞格式會影響結果的任務時，請保留具名的 V3.2 版本資訊。

進行實際的來源分析時，請為資料加上標籤，並要求結論由相關段落支持，同時區分假設與缺漏資訊。代理任務應說明允許的函式、預期最終結果與完成標準。這能讓模型的推理與工具工作流程圍繞可審閱的文字、程式碼或分析成果展開，並與所提供的證據連結。

", "aboutText": "DeepSeek V3.2 是 DeepSeek 用於推理與代理工作的文字模型，著重高效率的長上下文處理。製造商將主要變更歸納為 DeepSeek Sparse Attention、擴展規模的強化學習框架，以及代理任務的訓練管線。它接受書面問題與上下文，並產生文字，包括答案、程式碼或工具呼叫內容。它與較早的 V3.2 Exp 版本及獨立的 Speciale 推理變體有所區別。\n\nDeepSeek Sparse Attention 在潛在注意力架構中加入索引器與細粒度 token 選取。對長序列而言，索引器會選取主要注意力計算所用的先前 token 資訊。報告指出，這是相較於 V3.1 Terminus 的架構變更，且採用與 V3.2 Exp 相同的架構。公布的上下文長度為 128K token，可支援大量來源資料與持續進行的任務，但生成的推理與工具歷史也會占用可用上下文。\n\n後訓練涵蓋數學、程式設計、邏輯推理、一般代理任務、代理式程式撰寫與搜尋中的思考及非思考回應。各領域專家提供資料，彙整至最終模型後，再進行混合強化學習。一般問題可能適合直接回答；困難問題則可透過思考工作流程，在呈現最終結果前發展解法。請將所需答案或成果與中間推理分開指定。\n\n此版本的一項重大變更是結合工具進行思考。更新後的聊天格式讓推理能持續貫穿工具呼叫任務，而非將分析與操作視為無關的階段。模型選擇的操作由應用程式執行，再回傳結果供下一步使用。依製造商的工具上下文規則，只新增工具相關訊息時，會保留歷史推理；收到新的使用者訊息時，則會捨棄該推理，但保留工具呼叫與結果。正確區分對話角色會影響這項行為。\n\n程式撰寫任務應提供原始碼片段或相關檔案、觀察到的問題、預期結果，以及修改必須保留的條件。任務說明範例為：「使用提供的程式碼與工具結果找出原因，提出最小修正，並回傳最終修補內容，以及支持該修正的驗證結果。」搜尋任務應說明問題、所需證據，以及何種結果才算完成回答。搜尋與執行工具由周邊應用程式提供，並非單靠文字即可取得。\n\n製造商建議本機使用時將 temperature 設為 1.0、top-p 設為 0.95。V3.2 引入修訂的工具呼叫表示方式與專用訊息編碼器。其 developer 角色保留給製造商的搜尋代理範本，官方 API 不接受該角色。參考解析器要求格式正確的輸出，不會修復格式錯誤的結果；應用程式使用生成欄位或工具引數時，應先檢查結構再據此操作。\n\n技術報告指出，過長的推理歷程與重複自我驗證是實際限制。即使原始來源資料能容納於上下文中，工具輸出與重複推理仍會累積，使代理任務達到 128K 邊界。請精簡記錄相關觀察，並明確指定最終成果。對知識準確性要求較高的工作，請附上權威資料或提供適當的搜尋工具，而非期待訓練知識能確定每項事實。\n\nDeepSeek V3.2 Speciale 是獨立的高運算量深度推理變體；依模型卡所述，它不支援工具呼叫。不應以其功能描述標準 V3.2 的工具組態。同樣地，後來的 V4 世代與實驗性前代也各有不同身分與設定選項。比較模型與提示詞格式會影響結果的任務時，請保留具名的 V3.2 版本資訊。\n\n進行實際的來源分析時，請為資料加上標籤，並要求結論由相關段落支持，同時區分假設與缺漏資訊。代理任務應說明允許的函式、預期最終結果與完成標準。這能讓模型的推理與工具工作流程圍繞可審閱的文字、程式碼或分析成果展開，並與所提供的證據連結。", "editorialOverlay": { "copy_fields": null, "copy_sha256": null, "copy_source": null, "html_sha256": "ecb3f797925c814e1c57f3bd37903800abeb1af232877068cb936787e523d7cf", "module_sha256": "ee9dc1c50b1e6a8fb7fa062030fcb4646b9aaaeb08be513b723a5a92a7290f06", "raw_model_unchanged": true, "schema": "ailoot-native-model-editorial-v1" }, "locale": "zh-hant-tw", "methods": [ { "html": "

DS

DeepSeek: DeepSeek V3.2

DeepSeek /deepseek-v3.2-thinking

入門

DeepSeek V3.2 結合文字推理、工具使用與長上下文分析。其稀疏注意力架構與更新的對話格式支援程式撰寫、搜尋及多步驟代理任務，提供 128K 上下文，並區分思考與直接回應工作流程。

接受的輸入: 文字
結果: 文字

關於模型

此說明介紹模型的能力。Ailoot 可用的輸入、輸出與選項，依所選方法列於下方。

DeepSeek V3.2 是 DeepSeek 用於推理與代理工作的文字模型，著重高效率的長上下文處理。製造商將主要變更歸納為 DeepSeek Sparse Attention、擴展規模的強化學習框架，以及代理任務的訓練管線。它接受書面問題與上下文，並產生文字，包括答案、程式碼或工具呼叫內容。它與較早的 V3.2 Exp 版本及獨立的 Speciale 推理變體有所區別。

DeepSeek Sparse Attention 在潛在注意力架構中加入索引器與細粒度 token 選取。對長序列而言，索引器會選取主要注意力計算所用的先前 token 資訊。報告指出，這是相較於 V3.1 Terminus 的架構變更，且採用與 V3.2 Exp 相同的架構。公布的上下文長度為 128K token，可支援大量來源資料與持續進行的任務，但生成的推理與工具歷史也會占用可用上下文。

後訓練涵蓋數學、程式設計、邏輯推理、一般代理任務、代理式程式撰寫與搜尋中的思考及非思考回應。各領域專家提供資料，彙整至最終模型後，再進行混合強化學習。一般問題可能適合直接回答；困難問題則可透過思考工作流程，在呈現最終結果前發展解法。請將所需答案或成果與中間推理分開指定。

此版本的一項重大變更是結合工具進行思考。更新後的聊天格式讓推理能持續貫穿工具呼叫任務，而非將分析與操作視為無關的階段。模型選擇的操作由應用程式執行，再回傳結果供下一步使用。依製造商的工具上下文規則，只新增工具相關訊息時，會保留歷史推理；收到新的使用者訊息時，則會捨棄該推理，但保留工具呼叫與結果。正確區分對話角色會影響這項行為。

程式撰寫任務應提供原始碼片段或相關檔案、觀察到的問題、預期結果，以及修改必須保留的條件。任務說明範例為：「使用提供的程式碼與工具結果找出原因，提出最小修正，並回傳最終修補內容，以及支持該修正的驗證結果。」搜尋任務應說明問題、所需證據，以及何種結果才算完成回答。搜尋與執行工具由周邊應用程式提供，並非單靠文字即可取得。

製造商建議本機使用時將 temperature 設為 1.0、top-p 設為 0.95。V3.2 引入修訂的工具呼叫表示方式與專用訊息編碼器。其 developer 角色保留給製造商的搜尋代理範本，官方 API 不接受該角色。參考解析器要求格式正確的輸出，不會修復格式錯誤的結果；應用程式使用生成欄位或工具引數時，應先檢查結構再據此操作。

技術報告指出，過長的推理歷程與重複自我驗證是實際限制。即使原始來源資料能容納於上下文中，工具輸出與重複推理仍會累積，使代理任務達到 128K 邊界。請精簡記錄相關觀察，並明確指定最終成果。對知識準確性要求較高的工作，請附上權威資料或提供適當的搜尋工具，而非期待訓練知識能確定每項事實。

DeepSeek V3.2 Speciale 是獨立的高運算量深度推理變體；依模型卡所述，它不支援工具呼叫。不應以其功能描述標準 V3.2 的工具組態。同樣地，後來的 V4 世代與實驗性前代也各有不同身分與設定選項。比較模型與提示詞格式會影響結果的任務時，請保留具名的 V3.2 版本資訊。

進行實際的來源分析時，請為資料加上標籤，並要求結論由相關段落支持，同時區分假設與缺漏資訊。代理任務應說明允許的函式、預期最終結果與完成標準。這能讓模型的推理與工具工作流程圍繞可審閱的文字、程式碼或分析成果展開，並與所提供的證據連結。

Ailoot 可用 · 輸入與方法

Ailoot Ultra · OpenAI Chat POST /v1/chat/completions

輸入: text
輸出: text
輸出格式: json 或 SSE

欄位	類型	要求	值與限制	預設值
欄位`model`	類型`string`	要求必填	值與限制 deepseek-v3.2-thinking	預設值—
欄位`messages`	類型`message[]`	要求必填	值與限制項目數下限: 1	預設值—
欄位`messages[].role`	類型`string`	要求必填	值與限制依模型而定	預設值—
欄位`messages[].content`	類型`string`	要求選填	值與限制適用條件: 文字對話	預設值—
欄位`stream`	類型`boolean`	要求選填	值與限制	預設值false

輸入素材組合

文字 · UTF8 文字

messages[].content

方法參數與生命週期

參數	規格
參數`content_type`	規格application/json
參數`required`	規格model, messages
參數`material_types`	規格text
參數`messages`	規格具有 role/content 結構的訊息陣列
參數`file_material`	規格實際 UTF-8 文件以原文文字插入；不推定檔案 ID
參數`optional_controls`	規格stream
參數`control_scope`	規格文字對話；stream 是布林值，預設為 false。
參數`output_modalities`	規格text
參數`hosted_operations`	規格false
參數`multimodal_input_binding`	規格null

方法別名： /chat/completions

生命週期： synchronous_or_streamed

將確切模型 ID 與 messages 一併送出。

上下文長度不是輸出上限。

文字檔案可提供原文文字；圖片／音訊／檔案物件輸入需要精確綁定。

方案

費率資訊不會增加請求選項。支援的參數列於所選方法中。

快速模式: 未指定選取快速模式時
flex 模式: 未指定選取 flex 模式時
最高推理強度: 未指定最高推理強度採用 default 調整值時

Ailoot 費率

各設定的價格

設定	費率表	費用
設定詞元基本費率	費率表基本費率	費用輸入 · 輸入 Token $0.00058 / 1K Token 輸出 · 輸出 Token $0.00086 / 1K Token 輸入 · 讀取快取 $0 / 1K Token 輸入 · 寫入快取 $0 / 1K Token 輸入 · 寫入快取，1 小時 — / 1K Token

Ailoot 會在請求前預留估算金額。取得結果後，依實測用量結算最終費用。串流結算採用最終用量紀錄。預留金額與結算的運作方式 →

請求範例

Ailoot Ultra · OpenAI Chat

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-v3.2-thinking","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\n

from 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-v3.2-thinking", 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"))\n

接收並儲存結果

Ailoot Ultra · OpenAI Chat

json 或 SSE

保留所有已完成的結果

請保留所有完整傳回的回應、檔案與內容區塊。

保留每個候選項及其內容／工具呼叫／推理。SSE 中依候選項／工具索引組合增量，保留最終完成原因與用量。

需要檢查原始回應時

保留請求／回應 ID 與標頭。未提供通用的文字工作擷取 URL。

回應欄位

欄位	類型
欄位`choices[].index`	類型integer
欄位`choices[].message.content`	類型string\|null
欄位`choices[].message.tool_calls`	類型tool_call[]
欄位`choices[].message.reasoning_content`	類型string
欄位`choices[].message.reasoning_details`	類型object\|array
欄位`choices[].finish_reason`	類型string\|null
欄位`usage`	類型object

錯誤與後續步驟

回應	條件	下一步
回應400 · `invalid_request_error`	條件必要欄位、素材表示方式或參數值無效	下一步提出新的建立請求前，先修正指定欄位。

別名

模型別名

Ailoot Ultra deepseek-v3.2-thinking

方法路徑別名

/chat/completions

相關呼叫用 ID

deepseek-v3.2-thinking

", "method_identity": { "effective_request_model_id": "deepseek-v3.2-thinking", "fmt": "chat", "http": "POST", "path": "/v1/chat/completions", "request_model_id": "deepseek-v3.2-thinking", "route_label": "Ultra" }, "method_key": "Ultra|chat|/v1/chat/completions|deepseek-v3.2-thinking", "ordinal": 0, "rawMethod": { "available": null, "cache_read": 0, "cache_write": 0, "cache_write_1h": null, "contract": { "data": { "all_results": "Keep every choice and its content/tool calls/reasoning. In SSE assemble deltas by choice/tool index, retain the final finish reason and usage.", "authentication": { "api_key_source": "AILOOT_API_KEY", "header": "Authorization: Bearer" }, "errors": [ { "code": null, "condition": "必要欄位、素材表示方式或參數值無效", "http_status": 400, "recovery": "提出新的建立請求前，先修正指定欄位。", "type": "invalid_request_error" } ], "hints": [ "Model context is not an output cap.", "Send the conversation as text messages. Stream must be a boolean; omitted stream defaults to false." ], "ignored_fields": [], "lifecycle": { "create": { "http": "POST", "path": "/v1/chat/completions" }, "job_read": null, "mode": "synchronous_or_streamed", "unknown": "保留請求／回應 ID 與標頭。未提供通用的文字工作擷取 URL。" }, "material_combinations": [ { "mode": "text", "representation": "messages[].content", "required_material": "UTF8 文字" } ], "operation_modes": [ "conversation" ], "outputs": { "all_results": "保留每個候選項及其內容／工具呼叫／推理。SSE 中依候選項／工具索引組合增量，保留最終完成原因與用量。", "family": "text", "fields": [ { "description": null, "name": "choices[].index", "type": "integer" }, { "description": null, "name": "choices[].message.content", "type": "string|null" }, { "description": null, "name": "choices[].message.tool_calls", "type": "tool_call[]" }, { "description": null, "name": "choices[].message.reasoning_content", "type": "string" }, { "description": null, "name": "choices[].message.reasoning_details", "type": "object|array" }, { "description": null, "name": "choices[].finish_reason", "type": "string|null" }, { "description": null, "name": "usage", "type": "object" } ], "format": "json 或 SSE" }, "parameters": [ { "availability": "source_interface", "default": null, "default_known": false, "enum": [ "deepseek-v3.2-thinking" ], "enum_known": true, "name": "model", "required": true, "type": "string" }, { "availability": "source_interface", "default": null, "default_known": false, "enum": null, "enum_known": false, "min_items": 1, "name": "messages", "required": true, "type": "message[]" }, { "availability": "model_dependent", "default": null, "default_known": false, "enum": null, "enum_known": false, "name": "messages[].role", "required": true, "type": "string" }, { "availability": "source_interface", "default": null, "default_known": false, "enum": null, "enum_known": false, "name": "messages[].content", "required": false, "type": "string", "when": "文字對話" }, { "availability": "source_interface", "default": false, "default_known": true, "enum": null, "enum_known": false, "name": "stream", "required": false, "type": "boolean" } ], "profile_binding": "Text conversation at the exact model ID and method path.", "unsupported": [ "additional models/models list", "plugins", "route", "hosted tools; ordinary client function tools remain separate", "non-text output modalities", "provider.options overrides", "one-hour prompt-cache TTL" ] }, "hints": [ "將確切模型 ID 與 messages 一併送出。", "上下文長度不是輸出上限。", "文字檔案可提供原文文字；圖片／音訊／檔案物件輸入需要精確綁定。" ], "input": { "content_type": "application/json", "control_scope": "文字對話；stream 是布林值，預設為 false。", "file_material": "實際 UTF-8 文件以原文文字插入；不推定檔案 ID", "hosted_operations": false, "material_types": [ "text" ], "messages": "具有 role/content 結構的訊息陣列", "multimodal_input_binding": null, "optional_controls": [ "stream" ], "output_modalities": [ "text" ], "required": [ "model", "messages" ] }, "lifecycle": { "mode": "synchronous_or_streamed", "unknown": "preserve original request id; no generic chat job retrieval path" }, "manufacturer_capability": null, "method": { "http": "POST", "path": "/v1/chat/completions" }, "method_aliases": [ "/chat/completions" ], "operation_modes": [ "text_chat" ], "output": { "content": "choices[].message.content or stream delta", "family": "text", "format": "json or supported SSE", "tool_calls": "retain actual returned function calls" } }, "ctx": null, "declared_pricing_profiles": [ { "dimensions": { "regime": "base" }, "label": "詞元基本費率", "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } ], "examples": [ { "code": "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-v3.2-thinking\",\"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\n", "label": "curl · retained exact creation", "language": "bash" }, { "code": "from 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-v3.2-thinking\", 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\"))\n", "label": "Python SDK · all choices", "language": "python" } ], "family": "openai", "fast_multiplier": null, "flex_multiplier": null, "fmt": "chat", "hints": [ "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." ], "http": "POST", "in": 0.58, "input": [ "text" ], "label": "OpenAI Chat", "max_out": null, "max_reasoning_effort_multiplier": null, "method_aliases": [ "/chat/completions" ], "out": 0.86, "output": [ "text" ], "path": "/v1/chat/completions", "price_conditions": [ { "label": "Fast mode", "value": "Not specified", "when": "When fast mode is selected" }, { "label": "Flex mode", "value": "Not specified", "when": "When flex mode is selected" }, { "label": "Maximum reasoning effort", "value": "Not specified", "when": "When maximum reasoning effort uses the default adjustment" } ], "pricing": { "billing_mode": "token", "image_input_price": null, "image_output_price": null, "per_request_price": null }, "pricing_profiles": [ { "dimensions": { "regime": "base" }, "label": "Base token rates", "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } ], "request_model_id": "deepseek-v3.2-thinking", "route_label": "Ultra", "status_reads": [], "tiers": [], "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } } ], "model_id": "deepseek-v3.2-thinking", "nativeProvenance": { "input_facts_canonical_sha256": "a523f305b2bb6c8d25a11861a23bd4530dc92401119f3f1833b5cfe6e0155900", "input_html_sha256": "9e1567de44adc569dbcda33ed5dcf1a5e41759495f8609967f6a0a481a122084", "private_audit_sha256": "84e5d14e42989e02122409bbc65fbd46f3a21e4c0d58bd9545994e74e9367106", "role": "historical-normalized-input-audit-hashes", "schema": "ailoot-native-model-provenance-v2" }, "page": { "canonical": "https://ailoot.ai/models/zh-hant-tw/deepseek-v3-2-thinking/", "description": "DeepSeek V3.2 結合文字推理、工具使用與長上下文分析。其稀疏注意力架構與更新的對話格式支援程式撰寫、搜尋及多步驟代理任務，提供 128K 上下文，並區分思考與直接回應工作流程。", "locale": "zh-hant-tw", "md_url": "https://ailoot.ai/models/zh-hant-tw/deepseek-v3-2-thinking/index.md", "title": "DeepSeek: DeepSeek V3.2 — Ailoot", "url": "https://ailoot.ai/models/zh-hant-tw/deepseek-v3-2-thinking/" }, "rawModel": { "about": "DeepSeek V3.2 是 DeepSeek 用於推理與代理工作的文字模型，著重高效率的長上下文處理。製造商將主要變更歸納為 DeepSeek Sparse Attention、擴展規模的強化學習框架，以及代理任務的訓練管線。它接受書面問題與上下文，並產生文字，包括答案、程式碼或工具呼叫內容。它與較早的 V3.2 Exp 版本及獨立的 Speciale 推理變體有所區別。\n\nDeepSeek Sparse Attention 在潛在注意力架構中加入索引器與細粒度 token 選取。對長序列而言，索引器會選取主要注意力計算所用的先前 token 資訊。報告指出，這是相較於 V3.1 Terminus 的架構變更，且採用與 V3.2 Exp 相同的架構。公布的上下文長度為 128K token，可支援大量來源資料與持續進行的任務，但生成的推理與工具歷史也會占用可用上下文。\n\n後訓練涵蓋數學、程式設計、邏輯推理、一般代理任務、代理式程式撰寫與搜尋中的思考及非思考回應。各領域專家提供資料，彙整至最終模型後，再進行混合強化學習。一般問題可能適合直接回答；困難問題則可透過思考工作流程，在呈現最終結果前發展解法。請將所需答案或成果與中間推理分開指定。\n\n此版本的一項重大變更是結合工具進行思考。更新後的聊天格式讓推理能持續貫穿工具呼叫任務，而非將分析與操作視為無關的階段。模型選擇的操作由應用程式執行，再回傳結果供下一步使用。依製造商的工具上下文規則，只新增工具相關訊息時，會保留歷史推理；收到新的使用者訊息時，則會捨棄該推理，但保留工具呼叫與結果。正確區分對話角色會影響這項行為。\n\n程式撰寫任務應提供原始碼片段或相關檔案、觀察到的問題、預期結果，以及修改必須保留的條件。任務說明範例為：「使用提供的程式碼與工具結果找出原因，提出最小修正，並回傳最終修補內容，以及支持該修正的驗證結果。」搜尋任務應說明問題、所需證據，以及何種結果才算完成回答。搜尋與執行工具由周邊應用程式提供，並非單靠文字即可取得。\n\n製造商建議本機使用時將 temperature 設為 1.0、top-p 設為 0.95。V3.2 引入修訂的工具呼叫表示方式與專用訊息編碼器。其 developer 角色保留給製造商的搜尋代理範本，官方 API 不接受該角色。參考解析器要求格式正確的輸出，不會修復格式錯誤的結果；應用程式使用生成欄位或工具引數時，應先檢查結構再據此操作。\n\n技術報告指出，過長的推理歷程與重複自我驗證是實際限制。即使原始來源資料能容納於上下文中，工具輸出與重複推理仍會累積，使代理任務達到 128K 邊界。請精簡記錄相關觀察，並明確指定最終成果。對知識準確性要求較高的工作，請附上權威資料或提供適當的搜尋工具，而非期待訓練知識能確定每項事實。\n\nDeepSeek V3.2 Speciale 是獨立的高運算量深度推理變體；依模型卡所述，它不支援工具呼叫。不應以其功能描述標準 V3.2 的工具組態。同樣地，後來的 V4 世代與實驗性前代也各有不同身分與設定選項。比較模型與提示詞格式會影響結果的任務時，請保留具名的 V3.2 版本資訊。\n\n進行實際的來源分析時，請為資料加上標籤，並要求結論由相關段落支持，同時區分假設與缺漏資訊。代理任務應說明允許的函式、預期最終結果與完成標準。這能讓模型的推理與工具工作流程圍繞可審閱的文字、程式碼或分析成果展開，並與所提供的證據連結。", "alias_of": null, "aliases": [], "author": "DeepSeek", "capabilities": { "context_length": 163840, "manufacturer_input": [ "text" ], "manufacturer_output": [ "text" ] }, "card_routes": [ { "label": "Ultra", "rates": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" } ], "summary": "完整的設定價格資訊請見詳細資料。" } ], "ctx": null, "formats": [ "chat" ], "id": "deepseek-v3.2-thinking", "input": [ "text" ], "kind": "text", "max_out": null, "methods": [ { "available": null, "cache_read": 0, "cache_write": 0, "cache_write_1h": null, "contract": { "data": { "all_results": "Keep every choice and its content/tool calls/reasoning. In SSE assemble deltas by choice/tool index, retain the final finish reason and usage.", "authentication": { "api_key_source": "AILOOT_API_KEY", "header": "Authorization: Bearer" }, "errors": [ { "code": null, "condition": "必要欄位、素材表示方式或參數值無效", "http_status": 400, "recovery": "提出新的建立請求前，先修正指定欄位。", "type": "invalid_request_error" } ], "hints": [ "Model context is not an output cap.", "Send the conversation as text messages. Stream must be a boolean; omitted stream defaults to false." ], "ignored_fields": [], "lifecycle": { "create": { "http": "POST", "path": "/v1/chat/completions" }, "job_read": null, "mode": "synchronous_or_streamed", "unknown": "保留請求／回應 ID 與標頭。未提供通用的文字工作擷取 URL。" }, "material_combinations": [ { "mode": "text", "representation": "messages[].content", "required_material": "UTF8 文字" } ], "operation_modes": [ "conversation" ], "outputs": { "all_results": "保留每個候選項及其內容／工具呼叫／推理。SSE 中依候選項／工具索引組合增量，保留最終完成原因與用量。", "family": "text", "fields": [ { "description": null, "name": "choices[].index", "type": "integer" }, { "description": null, "name": "choices[].message.content", "type": "string|null" }, { "description": null, "name": "choices[].message.tool_calls", "type": "tool_call[]" }, { "description": null, "name": "choices[].message.reasoning_content", "type": "string" }, { "description": null, "name": "choices[].message.reasoning_details", "type": "object|array" }, { "description": null, "name": "choices[].finish_reason", "type": "string|null" }, { "description": null, "name": "usage", "type": "object" } ], "format": "json 或 SSE" }, "parameters": [ { "availability": "source_interface", "default": null, "default_known": false, "enum": [ "deepseek-v3.2-thinking" ], "enum_known": true, "name": "model", "required": true, "type": "string" }, { "availability": "source_interface", "default": null, "default_known": false, "enum": null, "enum_known": false, "min_items": 1, "name": "messages", "required": true, "type": "message[]" }, { "availability": "model_dependent", "default": null, "default_known": false, "enum": null, "enum_known": false, "name": "messages[].role", "required": true, "type": "string" }, { "availability": "source_interface", "default": null, "default_known": false, "enum": null, "enum_known": false, "name": "messages[].content", "required": false, "type": "string", "when": "文字對話" }, { "availability": "source_interface", "default": false, "default_known": true, "enum": null, "enum_known": false, "name": "stream", "required": false, "type": "boolean" } ], "profile_binding": "Text conversation at the exact model ID and method path.", "unsupported": [ "additional models/models list", "plugins", "route", "hosted tools; ordinary client function tools remain separate", "non-text output modalities", "provider.options overrides", "one-hour prompt-cache TTL" ] }, "hints": [ "將確切模型 ID 與 messages 一併送出。", "上下文長度不是輸出上限。", "文字檔案可提供原文文字；圖片／音訊／檔案物件輸入需要精確綁定。" ], "input": { "content_type": "application/json", "control_scope": "文字對話；stream 是布林值，預設為 false。", "file_material": "實際 UTF-8 文件以原文文字插入；不推定檔案 ID", "hosted_operations": false, "material_types": [ "text" ], "messages": "具有 role/content 結構的訊息陣列", "multimodal_input_binding": null, "optional_controls": [ "stream" ], "output_modalities": [ "text" ], "required": [ "model", "messages" ] }, "lifecycle": { "mode": "synchronous_or_streamed", "unknown": "preserve original request id; no generic chat job retrieval path" }, "manufacturer_capability": null, "method": { "http": "POST", "path": "/v1/chat/completions" }, "method_aliases": [ "/chat/completions" ], "operation_modes": [ "text_chat" ], "output": { "content": "choices[].message.content or stream delta", "family": "text", "format": "json or supported SSE", "tool_calls": "retain actual returned function calls" } }, "ctx": null, "declared_pricing_profiles": [ { "dimensions": { "regime": "base" }, "label": "詞元基本費率", "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } ], "examples": [ { "code": "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-v3.2-thinking\",\"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\n", "label": "curl · retained exact creation", "language": "bash" }, { "code": "from 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-v3.2-thinking\", 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\"))\n", "label": "Python SDK · all choices", "language": "python" } ], "family": "openai", "fast_multiplier": null, "flex_multiplier": null, "fmt": "chat", "hints": [ "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." ], "http": "POST", "in": 0.58, "input": [ "text" ], "label": "OpenAI Chat", "max_out": null, "max_reasoning_effort_multiplier": null, "method_aliases": [ "/chat/completions" ], "out": 0.86, "output": [ "text" ], "path": "/v1/chat/completions", "price_conditions": [ { "label": "Fast mode", "value": "Not specified", "when": "When fast mode is selected" }, { "label": "Flex mode", "value": "Not specified", "when": "When flex mode is selected" }, { "label": "Maximum reasoning effort", "value": "Not specified", "when": "When maximum reasoning effort uses the default adjustment" } ], "pricing": { "billing_mode": "token", "image_input_price": null, "image_output_price": null, "per_request_price": null }, "pricing_profiles": [ { "dimensions": { "regime": "base" }, "label": "Base token rates", "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } ], "request_model_id": "deepseek-v3.2-thinking", "route_label": "Ultra", "status_reads": [], "tiers": [], "units": [ { "amount": "0.00058000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "input_price", "label": "Input tokens", "native_amount_usd_per_unit": "0.00000058", "native_quantity": 1, "native_unit": "input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "input_token" }, { "amount": "0.00086000", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "output_price", "label": "Output tokens", "native_amount_usd_per_unit": "0.00000086", "native_quantity": 1, "native_unit": "output_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "out", "unit": "output_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_read_price", "label": "Cache read", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "cached_input_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "cached_input_token" }, { "amount": "0", "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_price", "label": "Cache write", "native_amount_usd_per_unit": "0", "native_quantity": 1, "native_unit": "generic_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "generic_cache_creation_token" }, { "amount": null, "amount_unit": "usd_per_thousand_tokens", "billing_input": true, "estimate_only": false, "key": "cache_write_1h_price", "label": "Cache write, 1 hour", "native_amount_usd_per_unit": null, "native_quantity": 1, "native_unit": "one_hour_cache_creation_token", "public_unit_label": "/ 1K tokens", "quantity": 1000, "quantity_is_estimate": false, "rate_role": "typed_native_retail_rate", "side": "in", "unit": "one_hour_cache_creation_token" } ] } ], "model_aliases": [ { "canonical_model_id": "deepseek-v3.2-thinking", "request_model_id": "deepseek-v3.2-thinking", "route_label": "Ultra" } ], "name": "DeepSeek: DeepSeek V3.2", "price": { "in": 0.58, "in_label": null, "out": 0.86, "out_label": null, "scope_label": "Token rates", "sort_in": 0.58, "sort_out": 0.86, "varies": false }, "short": "DeepSeek V3.2 結合文字推理、工具使用與長上下文分析。其稀疏注意力架構與更新的對話格式支援程式撰寫、搜尋及多步驟代理任務，提供 128K 上下文，並區分思考與直接回應工作流程。" }, "renderer_schema": { "method_order": "Original DTO methods order; every exact route/format/path/caller survives.", "prices": "Original public DTO fields and original native display HTML; no flattened or normalized offers." }, "representation": "exact-public-dto-and-native-renderer-html", "representationNotes": { "current_reference": { "about_json_pointer": "/aboutHtml", "methods": [ { "contract_json_pointer": "/methods/0/rawMethod/contract", "html_json_pointer": "/methods/0/html", "method_ordinal": 0 } ], "role": "current-semantic-reference" }, "raw_input": { "examples_role": "historical-input-not-current-ready-examples", "json_pointers": [ "/rawModel", "/methods/0/rawMethod" ], "role": "historical-owner-input" }, "schema": "ailoot-native-model-representations-v1" }, "schema": "ailoot-static-model-facts-v1", "schema_version": 1, "shortText": "DeepSeek V3.2 結合文字推理、工具使用與長上下文分析。其稀疏注意力架構與更新的對話格式支援程式撰寫、搜尋及多步驟代理任務，提供 128K 上下文，並區分思考與直接回應工作流程。", "source": { "audit_sha256": "2e00086bcbfc28e64d0cebe97aa47c17d8d5618963027f8ffb9dc6123d7d25b1", "bridge_map_sha256": "1125275768b9636fc8461bc7da5061239a81ddfe508e2557843ca42c58ab0036", "raw_model_json_sha256": "d87afd174c612183d0d20f6dbd0b4b2c9289bfe919ca1d9839ad2a856d837e7a", "renderer_sha256": "13a3fc1b22d07359278e2249d630bfc9bb9e6416c71ba53def933fb5a903660b" }, "sourceRevision": "eae9e8b18c204e6da6e0c919361524ed579f8e96327ef357dbced7f49bd2190a" }