[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-image-service-v2-gpt-6-luna":3,"text-service-v2-gpt-6-luna":207},[4],{"id":5,"aliases":6,"types":7,"pricing_group":10,"pricing":11,"display_name":5,"model_capabilities":28,"model_capabilities_zh":104,"introduction":153,"introduction_zh":154,"description":155,"description_zh":156,"title_focus":157,"title_focus_zh":158,"keyword_extras":159,"keyword_extras_zh":162,"model_data":165},"gpt-6-luna",[5],[8,9],"openai","openai\u002Fresponses\u002Fv1","openai\u002Fgpt",{"input_tokens_tiers":12},{"tiers":13},[14,21],{"max_tokens":15,"pricing":16},272000,{"input_tokens":17,"output_tokens":18,"cache_read_tokens":19,"cache_write_tokens":20},"0.1","0.5","0.01","0.125",{"max_tokens":22,"pricing":23},0,{"input_tokens":24,"output_tokens":25,"cache_read_tokens":26,"cache_write_tokens":27},"0.2","0.75","0.02","0.25",{"groups":29,"highlights":92,"overall":96,"scenario":99,"summary":102,"weakness":103},[30,69,82],{"items":31,"name":68},[32,38,43,47,51,55,60,64],{"detail":33,"method":34,"name":35,"samples":36,"score":37},"","Mixed","Agent Collaboration",18,67,{"detail":33,"method":39,"name":40,"samples":41,"score":42},"Objective","Code Generation",12,62,{"detail":33,"method":34,"name":44,"samples":45,"score":46},"Code Comprehension & Debugging",6,80,{"detail":33,"method":34,"name":48,"samples":49,"score":50},"Creative Writing",22,75,{"detail":33,"method":34,"name":52,"samples":53,"score":54},"Information Retrieval \u002F RAG",9,83,{"detail":33,"method":56,"name":57,"samples":58,"score":59},"Subjective","Long-Context Handling",16,86,{"detail":33,"method":34,"name":61,"samples":62,"score":63},"Math & Science Reasoning",14,40,{"detail":33,"method":56,"name":65,"samples":66,"score":67},"Multilingual Capability",11,87,"Advanced Capabilities",{"items":70,"name":81},[71,75,78],{"detail":33,"method":34,"name":72,"samples":73,"score":74},"General Knowledge",10,91,{"detail":33,"method":56,"name":76,"samples":77,"score":50},"Text Comprehension",23,{"detail":33,"method":34,"name":79,"samples":58,"score":80},"Logical Reasoning",61,"General Foundations",{"items":83,"name":91},[84,88],{"detail":33,"method":34,"name":85,"samples":86,"score":87},"Hallucination Control",15,78,{"detail":33,"method":56,"name":89,"samples":36,"score":90},"Instruction Following",79,"Safety & Compliance",[93,94,95],{"name":72,"score":74},{"name":65,"score":67},{"name":57,"score":59},{"rating":97,"score":98},"良好",74,{"fit":100,"name":101},81,"Translation & Localization \u002F Customer Support \u002F Enterprise Knowledge Q&A","GPT-6 Luna scored 74 overall, performing well — a fast knowledge-oriented model. It stands out in general knowledge (91%), multilingual capability (87%) and long-context handling (86%), answering reliably across languages and long documents at low cost, which fits localization and support at volume. Comparatively, math and science reasoning (40%) is markedly weak — quantitative work is effectively a blind spot. High-volume knowledge Q&A is where the economics work.",{"name":61,"note":33,"score":63},{"groups":105,"highlights":144,"overall":148,"scenario":149,"summary":151,"weakness":152},[106,128,137],{"items":107,"name":127},[108,111,114,116,118,120,123,125],{"detail":33,"method":109,"name":110,"samples":36,"score":37},"主\u002F客观混合","智能体协同",{"detail":33,"method":112,"name":113,"samples":41,"score":42},"客观","代码生成",{"detail":33,"method":109,"name":115,"samples":45,"score":46},"代码理解与调试",{"detail":33,"method":109,"name":117,"samples":49,"score":50},"创意写作",{"detail":33,"method":109,"name":119,"samples":53,"score":54},"信息检索 \u002F RAG",{"detail":33,"method":121,"name":122,"samples":58,"score":59},"主观","长文本处理",{"detail":33,"method":109,"name":124,"samples":62,"score":63},"数学与科学推理",{"detail":33,"method":121,"name":126,"samples":66,"score":67},"多语言能力","专项进阶能力",{"items":129,"name":136},[130,132,134],{"detail":33,"method":109,"name":131,"samples":73,"score":74},"通识知识",{"detail":33,"method":121,"name":133,"samples":77,"score":50},"文本理解",{"detail":33,"method":109,"name":135,"samples":58,"score":80},"逻辑推理","通用基础能力",{"items":138,"name":143},[139,141],{"detail":33,"method":109,"name":140,"samples":86,"score":87},"幻觉控制",{"detail":33,"method":121,"name":142,"samples":36,"score":90},"指令遵循","安全与合规能力",[145,146,147],{"name":131,"score":74},{"name":126,"score":67},{"name":122,"score":59},{"rating":97,"score":98},{"fit":100,"name":150},"翻译与本地化 \u002F 智能客服 \u002F 企业知识库问答","GPT-6 Luna 综合得分 74，整体表现良好，是一款速度导向的知识型模型。通识知识（91%）、多语言能力（87%）与长文本处理（86%）表现突出，低成本下跨语言长文档回答可靠，适合大批量翻译与客服任务。相对而言，数学与科学推理（40%）明显不足，定量工作基本是盲区。大批量知识问答最能体现它的经济性。",{"name":124,"note":33,"score":63},"GPT-6 Luna is the speed-and-cost member of the GPT-6 family, with broad knowledge, multilingual capability and steady long-context handling — built for high-volume Q&A and localization.","GPT-6 Luna是GPT-6系列的速度与成本型模型，知识广博、多语言能力与长上下文处理稳定，适合高频问答和本地化大批量任务。","GPT-6 Luna delivers dependable knowledge answers across languages and long documents at low cost. Integrate in one click via the SilvaMux unified API.","GPT-6 Luna以低成本提供跨语言长文档的可靠知识回答。可通过 SilvaMux 统一 API 一键快速接入。","Fast Knowledge Model","快速知识模型",[160,161],"low cost","knowledge Q&A",[163,164],"低成本","知识问答",{"summary":166,"recentAvailability":179,"observedAt":33,"bars":180,"metrics":185},[167,171,175],{"label":168,"value":169,"unit":170},"modelsDetail.performance.latency","15.28","s",{"label":172,"value":173,"unit":174},"modelsDetail.performance.throughput","97.67K","tokens\u002Fmin",{"label":176,"value":177,"unit":178},"modelsDetail.performance.availability","99.8","%","99.8%",{"degraded":181,"down":182},[],[183,184],41,32,[186,189,192,196,200,203],{"label":187,"value":188},"modelsDetail.performance.completionRate","97%",{"label":190,"value":191},"modelsDetail.performance.cacheHitRate","98%",{"label":193,"value":194,"unit":195},"modelsDetail.performance.tokenUsage","2,528,556,434","tokens",{"label":197,"value":198,"unit":199},"modelsDetail.performance.avgRequests","48","modelsDetail.performance.requestsUnit",{"label":201,"value":202,"unit":174},"modelsDetail.performance.tpm","97,665",{"label":204,"value":205,"unit":206},"modelsDetail.performance.qpm","660","req\u002Fmin",{"ttft":208,"tps":211,"e2e":214,"requests":217,"successfulRequests":218,"successRate":219,"availability":220},{"p50":209,"p95":210},1004,1962,{"p50":212,"p95":213},76,84,{"p50":215,"p95":216},3143,9367,173680,170488,"98.16",[221,221,221,221,221,221,221,221,221,221,221,221,221,221,221,221,221,222,221,221,221,221,221,221],"up","degraded"]