Pareto 26.9

{"about":"### Pareto 家族的前一版本\n\nPareto 26.9 是 Unbiased 複合模型的獨立版本。在 Pareto 家族中，一次請求由數個先進模型與開放模型參與處理，再將結果整合成共同答案。使用者操作的是一個 Pareto 模型，不需手動挑選每個元件。開發商指出，基礎模型的組合可能改變。\n\n穩定版 26.10 推出後，26.9 仍被保留。其精確識別碼 `unbiased/pareto-26.9` 可在應用程式中明確選擇前一版本。對既有工作流程及比較工作而言，若重點是指明使用版本，而非自動切換至下一版，這便有意義。固定名稱本身不代表內部組合也會保持不變。\n\n### 處理提供的資料\n\n在 Ailoot，此版本接受文字並生成文字答案。其上下文視窗為 262,144 個 token，包含指示、歷史、原始資料及答案；模型中繼資料標示的答案容量最多為 131,072 個 token。26.10 的上下文更大，因此選擇 26.9 時，宜預先挑出真正與任務相關的文件及片段。\n\n工作情境可圍繞數份文字的比對建立。提供需求、解決方案與選擇標準，再要求依每個標準附上理由的比較表。長文件請指定分析目的：找出矛盾、撰寫有內容的摘要、整理給作者的問題，或準備更新結構。為原始片段命名，有助於將結論連回請求內的資料。\n\n### 技術對話\n\n分析程式問題時，加入相關程式碼及已在自身環境取得的執行結果。要求解釋觀察、提出修改，並列出可用於驗證結果的指標。下一則訊息再提供驗證結果。這樣能將文字對話連結至真實任務，同時不假設模型自行存取檔案或執行環境。\n\n準備新文字時，適合指定讀者、綱要及所需詳細程度。答案可為文件草稿、輸入資料摘要、差異清單或後續行動順序；具體格式由請求決定。\n\nPareto 26.9 提供 Messages、Responses、Chat Completions 及文字 generateContent 格式。這些是整合客戶端的方式。此連線以文字傳入資料；選擇請求格式不會額外加入影像、程式碼執行或外部系統存取。","alias_of":null,"aliases":[],"author":"Unbiased","canonical_model":"unbiased/pareto-26.9","ctx":null,"formats":["chat","gemini","messages","responses"],"id":"unbiased/pareto-26.9","input":["text"],"kind":"text","max_out":null,"methods":[{"cache_read":0.5,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"anthropic","fast_multiplier":1,"flex_multiplier":1,"fmt":"messages","http":"POST","img_in":null,"img_out":null,"in":5.0,"input":["text"],"label":"Anthropic Messages","max_out":null,"max_tool_calls":null,"out":15.0,"output":["text"],"path":"/v1/messages","request_model_id":"unbiased/pareto-26.9","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":0.5,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"responses","http":"POST","img_in":null,"img_out":null,"in":5.0,"input":["text"],"label":"OpenAI Responses","max_out":null,"max_tool_calls":null,"out":15.0,"output":["text"],"path":"/v1/responses","request_model_id":"unbiased/pareto-26.9","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":0.5,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"chat","http":"POST","img_in":null,"img_out":null,"in":5.0,"input":["text"],"label":"OpenAI Chat","max_out":null,"max_tool_calls":null,"out":15.0,"output":["text"],"path":"/v1/chat/completions","request_model_id":"unbiased/pareto-26.9","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":0.5,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"gemini","fast_multiplier":1,"flex_multiplier":1,"fmt":"gemini","http":"POST","img_in":null,"img_out":null,"in":5.0,"input":["text"],"label":"Gemini generateContent","max_out":null,"max_tool_calls":null,"out":15.0,"output":["text"],"path":"/v1beta/models/unbiased/pareto-26.9:generateContent","request_model_id":"unbiased/pareto-26.9","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]}],"model_aliases":[{"canonical_model_id":"unbiased/pareto-26.9","request_model_id":"unbiased/pareto-26.9","route_label":"Ultra"}],"name":"Pareto 26.9","price":{"in":5.0,"in_label":null,"out":15.0,"out_label":null,"sort_in":5.0,"sort_out":15.0,"varies":false},"public_id":"unbiased/pareto-26.9","short":"Unbiased 複合文字模型的前一版本，上下文為 262,144 個 token，供明確固定版本的工作流程使用。"}