[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-gemini-3.5-flash":3},[4],{"id":5,"aliases":6,"types":7,"pricing_group":9,"pricing":10,"display_name":5,"capabilities":15,"model_capabilities":20,"model_data":94,"introduction":139,"description":140,"title_focus":141,"keyword_extras":142},"gemini-3.5-flash",[5],[8],"openai","google\u002Fgemini",{"flat_tokens":11},{"input_tokens":12,"output_tokens":13,"cache_read_tokens":14,"cache_write_tokens":12},"0.3","2.5","0.075",{"context_window":16,"max_output":17,"modalities":18},1048576,65536,[19],"text",{"groups":21,"highlights":83,"overall":87,"scenario":90,"summary":92,"weakness":93},[22,60,74],{"items":23,"name":59},[24,30,35,40,44,48,52,56],{"detail":25,"method":26,"name":27,"samples":28,"score":29},"","主\u002F客观混合","智能体协同",7,51,{"detail":25,"method":31,"name":32,"samples":33,"score":34},"客观","代码生成",10,80,{"detail":25,"method":36,"name":37,"samples":38,"score":39},"主观","代码理解与调试",12,66,{"detail":25,"method":36,"name":41,"samples":42,"score":43},"创意写作",17,78,{"detail":25,"method":26,"name":45,"samples":46,"score":47},"信息检索 \u002F RAG",13,92,{"detail":25,"method":36,"name":49,"samples":50,"score":51},"长文本处理",3,47,{"detail":25,"method":31,"name":53,"samples":54,"score":55},"数学与科学推理",14,21,{"detail":25,"method":26,"name":57,"samples":33,"score":58},"多语言能力",74,"专项进阶能力",{"items":61,"name":73},[62,66,69],{"detail":25,"method":26,"name":63,"samples":64,"score":65},"通识知识",15,76,{"detail":25,"method":36,"name":67,"samples":68,"score":47},"文本理解",5,{"detail":25,"method":26,"name":70,"samples":71,"score":72},"逻辑推理",20,85,"通用基础能力",{"items":75,"name":82},[76,80],{"detail":25,"method":36,"name":77,"samples":78,"score":79},"幻觉控制",9,83,{"detail":25,"method":36,"name":81,"samples":54,"score":43},"指令遵循","安全与合规能力",[84,85,86],{"name":67,"score":47},{"name":45,"score":47},{"name":70,"score":72},{"rating":88,"score":89},"良好",71,{"fit":89,"name":91},"代码研发 \u002F 高准确率专业场景","综合能力 71 分（良好）。优势集中在 文本理解（92%）、信息检索 \u002F RAG（92%）、逻辑推理（85%）。相对短板为 数学与科学推理（21%）。适合 代码研发 \u002F 高准确率专业场景 等场景。在 数学与科学推理 类任务上建议结合人工复核或专项验证。",{"name":53,"note":25,"score":55},{"bars":95,"metrics":100,"recentAvailability":127,"summary":128},{"degraded":96,"down":98},[97],1,[99],0,[101,104,107,111,115,119,123],{"label":102,"value":103},"完成率","99%",{"label":105,"value":106},"缓存命中率","0.0%",{"label":108,"unit":109,"value":110},"累计评测耗时","s","9630.5",{"label":112,"unit":113,"value":114},"Token 消耗","tokens","240,662",{"label":116,"unit":117,"value":118},"平均请求次数","次","1.0",{"label":120,"unit":121,"value":122},"TPM","tokens\u002Fmin","1,499",{"label":124,"unit":125,"value":126},"QPM","req\u002Fmin","0.9","99.33%",[129,132,135],{"label":130,"unit":109,"value":131},"延迟","13.85",{"label":133,"unit":121,"value":134},"吞吐量","1.5K",{"label":136,"unit":137,"value":138},"可用率","%","99.33","Gemini 3.5 Flash是一款面向规模化应用的多模态模型，在快速响应之外强化了编程和智能体能力，适合实时助手、代码工作流和高并发内容处理。","Gemini 3.5 Flash兼顾速度、编程与智能体执行，适合高频多模态处理和持续工具调用任务。可通过 SilvaMux 统一 API 一键快速接入。","多模态智能体",[143,144],"快速响应","代码任务"]