[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-image-service-v2-gemini-3.1-flash-lite":3,"text-service-v2-gemini-3.1-flash-lite":197},[4],{"id":5,"aliases":6,"types":7,"pricing_group":9,"pricing":10,"display_name":5,"capabilities":15,"model_capabilities":20,"model_capabilities_zh":96,"introduction":145,"introduction_zh":146,"description":147,"description_zh":148,"title_focus":149,"title_focus_zh":150,"keyword_extras":151,"keyword_extras_zh":154,"model_data":157},"gemini-3.1-flash-lite",[5],[8],"openai","google\u002Fgemini",{"flat_tokens":11},{"input_tokens":12,"output_tokens":13,"cache_read_tokens":14,"cache_write_tokens":12},"0.25","1.5","0.025",{"context_window":16,"max_output":17,"modalities":18},1048576,65536,[19],"text",{"groups":21,"highlights":84,"overall":88,"scenario":91,"summary":94,"weakness":95},[22,61,75],{"items":23,"name":60},[24,30,35,39,44,48,52,56],{"detail":25,"method":26,"name":27,"samples":28,"score":29},"","Mixed","Agent Collaboration",18,54,{"detail":25,"method":31,"name":32,"samples":33,"score":34},"Objective","Code Generation",12,67,{"detail":25,"method":26,"name":36,"samples":37,"score":38},"Code Comprehension & Debugging",6,62,{"detail":25,"method":40,"name":41,"samples":42,"score":43},"Subjective","Creative Writing",22,57,{"detail":25,"method":26,"name":45,"samples":46,"score":47},"Information Retrieval \u002F RAG",9,53,{"detail":25,"method":40,"name":49,"samples":50,"score":51},"Long-Context Handling",16,73,{"detail":25,"method":26,"name":53,"samples":54,"score":55},"Math & Science Reasoning",14,77,{"detail":25,"method":40,"name":57,"samples":58,"score":59},"Multilingual Capability",11,81,"Advanced Capabilities",{"items":62,"name":74},[63,67,71],{"detail":25,"method":26,"name":64,"samples":65,"score":66},"General Knowledge",10,90,{"detail":25,"method":40,"name":68,"samples":69,"score":70},"Text Comprehension",23,80,{"detail":25,"method":26,"name":72,"samples":50,"score":73},"Logical Reasoning",65,"General Foundations",{"items":76,"name":83},[77,80],{"detail":25,"method":40,"name":78,"samples":79,"score":59},"Hallucination Control",15,{"detail":25,"method":40,"name":81,"samples":28,"score":82},"Instruction Following",71,"Safety & Compliance",[85,86,87],{"name":64,"score":66},{"name":57,"score":59},{"name":78,"score":59},{"rating":89,"score":90},"良好",70,{"fit":92,"name":93},75,"Translation & Localization \u002F Customer Support \u002F Content Moderation & Risk Control","Gemini 3.1 Flash Lite scored 70 overall, performing well — an ultra-lightweight model for cost-driven scale. It stands out in general knowledge (90%), multilingual capability (81%) and hallucination control (81%), delivering dependable answers at minimal cost for localization and support at volume. Comparatively, information retrieval \u002F RAG (53%) is weak; retrieval-grounded answers need verification. Lowest cost per unit of dependable multilingual output is its pitch.",{"name":45,"note":25,"score":47},{"groups":97,"highlights":136,"overall":140,"scenario":141,"summary":143,"weakness":144},[98,120,129],{"items":99,"name":119},[100,103,106,108,111,113,115,117],{"detail":25,"method":101,"name":102,"samples":28,"score":29},"主\u002F客观混合","智能体协同",{"detail":25,"method":104,"name":105,"samples":33,"score":34},"客观","代码生成",{"detail":25,"method":101,"name":107,"samples":37,"score":38},"代码理解与调试",{"detail":25,"method":109,"name":110,"samples":42,"score":43},"主观","创意写作",{"detail":25,"method":101,"name":112,"samples":46,"score":47},"信息检索 \u002F RAG",{"detail":25,"method":109,"name":114,"samples":50,"score":51},"长文本处理",{"detail":25,"method":101,"name":116,"samples":54,"score":55},"数学与科学推理",{"detail":25,"method":109,"name":118,"samples":58,"score":59},"多语言能力","专项进阶能力",{"items":121,"name":128},[122,124,126],{"detail":25,"method":101,"name":123,"samples":65,"score":66},"通识知识",{"detail":25,"method":109,"name":125,"samples":69,"score":70},"文本理解",{"detail":25,"method":101,"name":127,"samples":50,"score":73},"逻辑推理","通用基础能力",{"items":130,"name":135},[131,133],{"detail":25,"method":109,"name":132,"samples":79,"score":59},"幻觉控制",{"detail":25,"method":109,"name":134,"samples":28,"score":82},"指令遵循","安全与合规能力",[137,138,139],{"name":123,"score":66},{"name":118,"score":59},{"name":132,"score":59},{"rating":89,"score":90},{"fit":92,"name":142},"翻译与本地化 \u002F 智能客服 \u002F 内容审核与风控","Gemini 3.1 Flash Lite 综合得分 70，整体表现良好，是一款成本导向的超轻量模型。通识知识（90%）、多语言能力（81%）与幻觉控制（81%）表现突出，以极小成本提供可靠回答，适合大批量翻译与客服任务。相对而言，信息检索 \u002F RAG（53%）偏弱，检索增强类回答需外部校验。用最低成本获得可靠多语言输出是它的定位。",{"name":112,"note":25,"score":47},"Gemini 3.1 Flash Lite is Google's ultra-light model, offering steady knowledge and multilingual processing at minimal cost — built for high-volume localization, support and content moderation.","Gemini 3.1 Flash Lite是Google的超轻量模型，以极低成本提供稳定的通识知识与多语言处理能力，适合翻译本地化、智能客服和内容审核等规模化任务。","Gemini 3.1 Flash Lite focuses on low-cost, high-throughput serving for high-frequency lightweight Q&A and multilingual content. Integrate in one click via the SilvaMux unified API.","Gemini 3.1 Flash Lite主打低成本高吞吐，适合高频轻量问答与多语言内容处理。可通过 SilvaMux 统一 API 一键快速接入。","Lightweight Low-Cost Model","轻量低成本模型",[152,153],"low cost","high concurrency",[155,156],"低成本","高并发",{"summary":158,"recentAvailability":171,"observedAt":25,"bars":172,"metrics":175},[159,163,167],{"label":160,"value":161,"unit":162},"modelsDetail.performance.latency","22.60","s",{"label":164,"value":165,"unit":166},"modelsDetail.performance.throughput","96.40K","tokens\u002Fmin",{"label":168,"value":169,"unit":170},"modelsDetail.performance.availability","99.3","%","99.3%",{"degraded":173,"down":174},[],[],[176,179,182,186,190,193],{"label":177,"value":178},"modelsDetail.performance.completionRate","100%",{"label":180,"value":181},"modelsDetail.performance.cacheHitRate","91%",{"label":183,"value":184,"unit":185},"modelsDetail.performance.tokenUsage","1,646,843,697","tokens",{"label":187,"value":188,"unit":189},"modelsDetail.performance.avgRequests","35","modelsDetail.performance.requestsUnit",{"label":191,"value":192,"unit":166},"modelsDetail.performance.tpm","96,397",{"label":194,"value":195,"unit":196},"modelsDetail.performance.qpm","604","req\u002Fmin",{"ttft":198,"tps":201,"e2e":204,"requests":207,"successfulRequests":208,"successRate":209,"availability":210},{"p50":199,"p95":200},2443,2447,{"p50":202,"p95":203},32,82,{"p50":205,"p95":206},4928,9493,111631,110877,"99.32",[211,211,211,211,211,211,211,212,211,211,211,211,211,211,211,211,212,211,211,211,211,211,211,211],"up","degraded"]