[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-qwen3.6-plus":3},[4],{"id":5,"aliases":6,"types":7,"pricing":10,"capabilities":41,"model_capabilities":48,"model_data":119,"introduction":162,"description":163,"title_focus":164,"keyword_extras":165},"qwen3.6-plus",[5],[8,9],"openai","anthropic",{"input_tokens_tiers":11,"builtin_tool_calls":27},{"tiers":12},[13,20],{"max_tokens":14,"pricing":15},262144,{"input_tokens":16,"output_tokens":17,"cache_read_tokens":18,"cache_write_tokens":19},"2","12","0.2","2.5",{"max_tokens":21,"pricing":22},0,{"input_tokens":23,"output_tokens":24,"cache_read_tokens":25,"cache_write_tokens":26},"8","48","0.8","10",[28,31,34,36,39],{"tool_name":29,"price_per_unit":30},"web_search","4",{"tool_name":32,"price_per_unit":33},"web_extractor","0",{"tool_name":35,"price_per_unit":33},"code_interpreter",{"tool_name":37,"price_per_unit":38},"t2i_search","24",{"tool_name":40,"price_per_unit":24},"i2i_search",{"context_window":42,"max_output":43,"modalities":44},1000000,32768,[45,46,47],"text","image","video",{"groups":49,"highlights":108,"overall":112,"scenario":115,"summary":117,"weakness":118},[50,86,99],{"items":51,"name":85},[52,58,63,67,71,75,78,82],{"detail":53,"method":54,"name":55,"samples":56,"score":57},"","主\u002F客观混合","智能体协同",7,85,{"detail":53,"method":59,"name":60,"samples":61,"score":62},"客观","代码生成",10,80,{"detail":53,"method":54,"name":64,"samples":65,"score":66},"代码理解与调试",12,77,{"detail":53,"method":68,"name":69,"samples":70,"score":62},"主观","创意写作",17,{"detail":53,"method":54,"name":72,"samples":73,"score":74},"信息检索 \u002F RAG",13,87,{"detail":53,"method":68,"name":76,"samples":77,"score":57},"长文本处理",3,{"detail":53,"method":59,"name":79,"samples":80,"score":81},"数学与科学推理",14,93,{"detail":53,"method":54,"name":83,"samples":61,"score":84},"多语言能力",86,"专项进阶能力",{"items":87,"name":98},[88,91,94],{"detail":53,"method":54,"name":89,"samples":90,"score":62},"通识知识",15,{"detail":53,"method":68,"name":92,"samples":93,"score":81},"文本理解",5,{"detail":53,"method":54,"name":95,"samples":96,"score":97},"逻辑推理",20,73,"通用基础能力",{"items":100,"name":107},[101,104],{"detail":53,"method":68,"name":102,"samples":103,"score":74},"幻觉控制",9,{"detail":53,"method":68,"name":105,"samples":80,"score":106},"指令遵循",81,"安全与合规能力",[109,110,111],{"name":92,"score":81},{"name":79,"score":81},{"name":72,"score":74},{"rating":113,"score":114},"优秀",84,{"fit":114,"name":116},"翻译与本地化 \u002F 智能客服 \u002F 短剧剧本生成","Qwen3.6 Plus在本轮能力评测中综合得分84，整体表现优秀。其文本理解（93%）和数学与科学推理（93%）表现突出，能够分析文档语义、提取关键信息并处理复杂问答，并处理数学求解、科学分析和结构化推理任务。对企业知识库团队、AI 应用开发者和研究人员而言，它适合用于资料检索、文档问答、信息抽取和检索增强生成场景，解决从大量资料中找到相关信息，并形成可用于后续决策的回答；信息检索\u002FRAG（87%）也为相关任务提供补充。在能力边界上，逻辑推理（72%）相对较弱，在复杂多层推导任务中，建议核验中间过程而非只采信最终答案。",{"name":95,"note":53,"score":97},{"bars":120,"metrics":123,"recentAvailability":150,"summary":151},{"degraded":121,"down":122},[],[21],[124,127,130,134,138,142,146],{"label":125,"value":126},"完成率","96%",{"label":128,"value":129},"缓存命中率","动态变化",{"label":131,"unit":132,"value":133},"累计评测耗时","s","7544.4",{"label":135,"unit":136,"value":137},"Token 消耗","tokens","1,795,053",{"label":139,"unit":140,"value":141},"平均请求次数","次","9.2",{"label":143,"unit":144,"value":145},"TPM","tokens\u002Fmin","14,276",{"label":147,"unit":148,"value":149},"QPM","req\u002Fmin","1.8","99.5%",[152,155,158],{"label":153,"unit":132,"value":154},"延迟","8.97",{"label":156,"unit":144,"value":157},"吞吐量","14.28K",{"label":159,"unit":160,"value":161},"可用率","%","99.5","Qwen3.6 Plus是一款重视推理与知识获取的语言模型，能够完成数学科学分析、检索增强问答和多来源资料整合，适合研究、教育与企业知识场景。","Qwen3.6 Plus具备文本理解、数学与科学推理、信息检索\u002FRAG能力，可完成文档分析、信息抽取和复杂问答，并支持数学求解、科学分析和结构化推理。面向企业知识库团队、AI 应用开发者和研究人员，适合资料检索、文档问答、信息抽取和检索增强生成；可通过 SilvaMux 统一 API 一键快速接入。","数学检索模型",[166,167],"数学科学推理","企业知识问答"]