[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-image-service-v2-gemini-3.7-flash":3,"text-service-v2-gemini-3.7-flash":196},[4],{"id":5,"aliases":6,"types":7,"pricing_group":9,"pricing":10,"display_name":5,"capabilities":15,"model_capabilities":20,"model_capabilities_zh":94,"introduction":143,"introduction_zh":144,"description":145,"description_zh":146,"title_focus":147,"title_focus_zh":148,"keyword_extras":149,"keyword_extras_zh":152,"model_data":155},"gemini-3.7-flash",[5],[8],"openai","google\u002Fgemini",{"output_tokens":11,"flat_tokens":12},"7.5",{"input_tokens":13,"output_tokens":11,"cache_read_tokens":14,"cache_write_tokens":13},"1.5","0.15",{"context_window":16,"max_output":17,"modalities":18},1048576,65536,[19],"text",{"groups":21,"highlights":82,"overall":86,"scenario":89,"summary":92,"weakness":93},[22,60,73],{"items":23,"name":59},[24,30,35,39,44,47,51,55],{"detail":25,"method":26,"name":27,"samples":28,"score":29},"","Mixed","Agent Collaboration",18,73,{"detail":25,"method":31,"name":32,"samples":33,"score":34},"Objective","Code Generation",12,50,{"detail":25,"method":26,"name":36,"samples":37,"score":38},"Code Comprehension & Debugging",6,86,{"detail":25,"method":40,"name":41,"samples":42,"score":43},"Subjective","Creative Writing",22,77,{"detail":25,"method":26,"name":45,"samples":46,"score":43},"Information Retrieval \u002F RAG",9,{"detail":25,"method":26,"name":48,"samples":49,"score":50},"Long-Context Handling",16,79,{"detail":25,"method":26,"name":52,"samples":53,"score":54},"Math & Science Reasoning",14,83,{"detail":25,"method":40,"name":56,"samples":57,"score":58},"Multilingual Capability",11,91,"Advanced Capabilities",{"items":61,"name":72},[62,66,69],{"detail":25,"method":26,"name":63,"samples":64,"score":65},"General Knowledge",10,94,{"detail":25,"method":26,"name":67,"samples":68,"score":43},"Text Comprehension",23,{"detail":25,"method":26,"name":70,"samples":49,"score":71},"Logical Reasoning",67,"General Foundations",{"items":74,"name":81},[75,79],{"detail":25,"method":26,"name":76,"samples":77,"score":78},"Hallucination Control",15,76,{"detail":25,"method":26,"name":80,"samples":28,"score":50},"Instruction Following","Safety & Compliance",[83,84,85],{"name":63,"score":65},{"name":56,"score":58},{"name":36,"score":38},{"rating":87,"score":88},"良好",78,{"fit":90,"name":91},82,"Translation & Localization \u002F Customer Support \u002F Enterprise Knowledge Q&A","Gemini 3.7 Flash scored 78 overall, performing well — an efficient model with surprising depth. It stands out in general knowledge (94%), multilingual capability (91%) and code comprehension and debugging (86%), reading and fixing code credibly while handling multilingual knowledge work, which fits localization, support and knowledge-base Q&A at volume. Comparatively, code generation (50%) is weak; writing code from scratch should go to stronger models. Fast multilingual knowledge work with light debugging is its sweet spot.",{"name":32,"note":25,"score":34},{"groups":95,"highlights":134,"overall":138,"scenario":139,"summary":141,"weakness":142},[96,118,127],{"items":97,"name":117},[98,101,104,106,109,111,113,115],{"detail":25,"method":99,"name":100,"samples":28,"score":29},"主\u002F客观混合","智能体协同",{"detail":25,"method":102,"name":103,"samples":33,"score":34},"客观","代码生成",{"detail":25,"method":99,"name":105,"samples":37,"score":38},"代码理解与调试",{"detail":25,"method":107,"name":108,"samples":42,"score":43},"主观","创意写作",{"detail":25,"method":99,"name":110,"samples":46,"score":43},"信息检索 \u002F RAG",{"detail":25,"method":99,"name":112,"samples":49,"score":50},"长文本处理",{"detail":25,"method":99,"name":114,"samples":53,"score":54},"数学与科学推理",{"detail":25,"method":107,"name":116,"samples":57,"score":58},"多语言能力","专项进阶能力",{"items":119,"name":126},[120,122,124],{"detail":25,"method":99,"name":121,"samples":64,"score":65},"通识知识",{"detail":25,"method":99,"name":123,"samples":68,"score":43},"文本理解",{"detail":25,"method":99,"name":125,"samples":49,"score":71},"逻辑推理","通用基础能力",{"items":128,"name":133},[129,131],{"detail":25,"method":99,"name":130,"samples":77,"score":78},"幻觉控制",{"detail":25,"method":99,"name":132,"samples":28,"score":50},"指令遵循","安全与合规能力",[135,136,137],{"name":121,"score":65},{"name":116,"score":58},{"name":105,"score":38},{"rating":87,"score":88},{"fit":90,"name":140},"翻译与本地化 \u002F 智能客服 \u002F 企业知识库问答","Gemini 3.7 Flash 综合得分 78，整体表现良好，是一款深度超出预期的高效模型。通识知识（94%）、多语言能力（91%）与代码理解与调试（86%）表现突出，能可信地读懂并修复代码，同时胜任多语言知识工作，在翻译本地化与知识库问答场景中优势明显。相对而言，代码生成（50%）偏弱，从零写码建议交给更强的模型。快速多语言知识工作加轻量调试是它的最佳场景。",{"name":103,"note":25,"score":34},"Gemini 3.7 Flash is Google's efficient model in the 3.x line, combining broad knowledge and multilingual strength with credible code reading — suited to localization, support and enterprise knowledge bases.","Gemini 3.7 Flash是Google 3.x系列的高效模型，兼具广博知识、多语言能力与可信的代码阅读能力，适合翻译本地化、智能客服和企业知识库场景。","Gemini 3.7 Flash balances speed with knowledge, multilingual capability and code comprehension for high-volume daily work. Integrate in one click via the SilvaMux unified API.","Gemini 3.7 Flash平衡速度与知识、多语言和代码理解能力，适合高频日常工作。可通过 SilvaMux 统一 API 一键快速接入。","Efficient Knowledge Model","高效知识模型",[150,151],"fast response","multilingual",[153,154],"快速响应","多语言",{"summary":156,"recentAvailability":169,"observedAt":25,"bars":170,"metrics":174},[157,161,165],{"label":158,"value":159,"unit":160},"modelsDetail.performance.latency","49.35","s",{"label":162,"value":163,"unit":164},"modelsDetail.performance.throughput","92.63K","tokens\u002Fmin",{"label":166,"value":167,"unit":168},"modelsDetail.performance.availability","98.9","%","98.9%",{"degraded":171,"down":172},[57],[46,173],61,[175,178,181,185,189,192],{"label":176,"value":177},"modelsDetail.performance.completionRate","99%",{"label":179,"value":180},"modelsDetail.performance.cacheHitRate","96%",{"label":182,"value":183,"unit":184},"modelsDetail.performance.tokenUsage","2,888,911,466","tokens",{"label":186,"value":187,"unit":188},"modelsDetail.performance.avgRequests","42","modelsDetail.performance.requestsUnit",{"label":190,"value":191,"unit":164},"modelsDetail.performance.tpm","92,634",{"label":193,"value":194,"unit":195},"modelsDetail.performance.qpm","768","req\u002Fmin",{"ttft":197,"tps":200,"e2e":203,"requests":206,"successfulRequests":207,"successRate":208,"availability":209},{"p50":198,"p95":199},1776,2173,{"p50":201,"p95":202},34,89,{"p50":204,"p95":205},7647,9568,152577,152109,"99.69",[210,210,210,210,210,210,210,210,210,210,210,210,210,210,210,210,210,210,210,211,210,210,210,210],"up","degraded"]