[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-gpt-5.6-luna":3},[4],{"id":5,"aliases":6,"types":7,"pricing":10,"priority":27,"model_capabilities":28,"model_data":101,"introduction":144,"description":145,"title_focus":146,"keyword_extras":147},"gpt-5.6-luna",[5],[8,9],"openai","openai\u002Fresponses\u002Fv1",{"input_tokens_tiers":11},{"tiers":12},[13,20],{"max_tokens":14,"pricing":15},272000,{"input_tokens":16,"output_tokens":17,"cache_read_tokens":18,"cache_write_tokens":19},"0.2","1.2","0.02","0.25",{"max_tokens":21,"pricing":22},0,{"input_tokens":23,"output_tokens":24,"cache_read_tokens":25,"cache_write_tokens":26},"0.4","1.8","0.04","0.5",1,{"groups":29,"highlights":91,"overall":95,"scenario":97,"summary":99,"weakness":100},[30,68,82],{"items":31,"name":67},[32,38,43,48,52,56,60,64],{"detail":33,"method":34,"name":35,"samples":36,"score":37},"","主\u002F客观混合","智能体协同",7,56,{"detail":33,"method":39,"name":40,"samples":41,"score":42},"客观","代码生成",10,100,{"detail":33,"method":44,"name":45,"samples":46,"score":47},"主观","代码理解与调试",12,85,{"detail":33,"method":44,"name":49,"samples":50,"score":51},"创意写作",17,81,{"detail":33,"method":34,"name":53,"samples":54,"score":55},"信息检索 \u002F RAG",13,94,{"detail":33,"method":39,"name":57,"samples":58,"score":59},"数学与科学推理",14,86,{"detail":33,"method":44,"name":61,"samples":62,"score":63},"长文本处理",3,98,{"detail":33,"method":34,"name":65,"samples":41,"score":66},"多语言能力",87,"专项进阶能力",{"items":69,"name":81},[70,73,77],{"detail":33,"method":34,"name":71,"samples":72,"score":51},"通识知识",15,{"detail":33,"method":44,"name":74,"samples":75,"score":76},"文本理解",5,97,{"detail":33,"method":34,"name":78,"samples":79,"score":80},"逻辑推理",20,76,"通用基础能力",{"items":83,"name":90},[84,87],{"detail":33,"method":44,"name":85,"samples":86,"score":51},"幻觉控制",9,{"detail":33,"method":44,"name":88,"samples":58,"score":89},"指令遵循",80,"安全与合规能力",[92,93,94],{"name":40,"score":42},{"name":61,"score":63},{"name":74,"score":76},{"rating":96,"score":47},"优秀",{"fit":47,"name":98},"代码研发 \u002F 内容创作","综合能力 85 分（优秀）。优势集中在 代码生成（100%）、长文本处理（98%）、文本理解（97%）。相对短板为 智能体协同（56%）。适合 代码研发 \u002F 内容创作 等场景。",{"name":35,"note":33,"score":37},{"bars":102,"metrics":105,"recentAvailability":132,"summary":133},{"degraded":103,"down":104},[],[],[106,109,112,116,120,124,128],{"label":107,"value":108},"完成率","100%",{"label":110,"value":111},"缓存命中率","0.0%",{"label":113,"unit":114,"value":115},"累计评测耗时","s","1811.4",{"label":117,"unit":118,"value":119},"Token 消耗","tokens","168,840",{"label":121,"unit":122,"value":123},"平均请求次数","次","1.0",{"label":125,"unit":126,"value":127},"TPM","tokens\u002Fmin","5,592",{"label":129,"unit":130,"value":131},"QPM","req\u002Fmin","4.9","100.0%",[134,137,140],{"label":135,"unit":114,"value":136},"延迟","7.68",{"label":138,"unit":126,"value":139},"吞吐量","5.59K",{"label":141,"unit":142,"value":143},"可用率","%","100.0","GPT-5.6 Luna是GPT-5.6系列中偏重速度与成本效率的模型，适合批量分类、信息抽取、客服问答和延迟敏感的轻量工作流。","GPT-5.6 Luna强调快速和高性价比响应，适合高频问答、信息抽取与轻量自动化任务。可通过 SilvaMux 统一 API 一键快速接入。","快速响应模型",[148,149],"高频问答","信息抽取"]