NVIDIA: Nemotron 3.5 Super VL (free)

{"about":"### 推理、程式碼與長篇文字\n\nNemotron 3.5 Super VL 是 NVIDIA 用於多步驟推理、程式碼理解或大量資料處理的模型。在 Ailoot，本模型卡提供文字對話：傳入指示、輸入資料與訊息歷史，模型再傳回文字答案。\n\n可用於分析技術文件、比對需求、準備程式碼說明及分析文字報告。請求可先指定目標、附上必要資料，再定義結果形式，例如附理由的結論清單、程式修改計畫，或數個方案的比較。\n\n### 結構與專長\n\n原始 Super VL 是 NVIDIA 採用 C-RADIO 視覺編碼器的視覺語言模型，也針對文件與影片理解設計。這張 Ailoot 模型卡提供文字處理能力，原始資料以文字內容傳入。\n\n模型基礎結合 Mamba-2、Transformer 與 Mixture of Experts 架構。約 1200 億參數中，每個處理步驟啟用約 120 億參數。這是模型結構，應與個別請求的實際速度分開考量。\n\n訓練使用了程式儲存庫、任務與程式碼修改配對、建置日誌、多步驟解題軌跡及長上下文資料。後續訓練結合程式碼、推理、指示與對話。面對需要先連結輸入資訊、再整理結果的複合任務時，可運用這種專長來表達需求。\n\n### 如何準備任務\n\n文件分析先說明問題與資訊篩選標準，再附上有章節標題的文字。需要比對來源時，請給予簡短標籤，並要求結論對應至標籤。程式碼工作則加入相關程式片段、觀察到的行為與預期結果，讓答案能與具體情況核對。\n\n大型上下文視窗可在同一請求放入相關資料。容量本身不能取代清楚提問：適合將指示與原始資料分開，並預先指出所需方案或文件。\n\n### 上下文與連線方式\n\nAiloot 為此版本標示 1000000 個 token 的視窗，答案上限為 131072 個 token。對話歷史、指示、原始資料與答案共用上下文；準備長請求時，請保留結果所需容量。\n\n模型可透過 Chat Completions、Responses 與 Messages 使用。此版本提供文字輸入與文字輸出，文件以文字內容傳入。","alias_of":null,"aliases":[],"author":"NVIDIA","canonical_model":"nvidia/nemotron-3.5-super-vl:free","ctx":1000000,"formats":["chat","messages","responses"],"id":"nvidia/nemotron-3.5-super-vl:free","input":["text"],"kind":"text","max_out":131072,"methods":[{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":1000000,"family":"anthropic","fast_multiplier":1,"flex_multiplier":1,"fmt":"messages","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"Anthropic Messages","max_out":131072,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/messages","request_model_id":"nvidia/nemotron-3.5-super-vl:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":1000000,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"responses","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"OpenAI Responses","max_out":131072,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/responses","request_model_id":"nvidia/nemotron-3.5-super-vl:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":1000000,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"chat","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"OpenAI Chat","max_out":131072,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/chat/completions","request_model_id":"nvidia/nemotron-3.5-super-vl:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]}],"model_aliases":[{"canonical_model_id":"nvidia/nemotron-3.5-super-vl:free","request_model_id":"nvidia/nemotron-3.5-super-vl:free","route_label":"Ultra"}],"name":"NVIDIA: Nemotron 3.5 Super VL (free)","price":{"in":0,"in_label":null,"out":0,"out_label":null,"sort_in":0,"sort_out":0,"varies":false},"public_id":"nvidia/nemotron-3.5-super-vl:free","short":"與 NVIDIA 模型進行文字對話，用於推理、程式碼工作與大量資料分析。"}