[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-image-service-v2-claude-haiku-5-5":3,"text-service-v2-claude-haiku-5-5":299},[4,201],{"id":5,"aliases":6,"types":7,"pricing_group":10,"pricing":11,"display_name":5,"model_capabilities":27,"model_capabilities_zh":100,"introduction":149,"introduction_zh":150,"description":151,"description_zh":152,"title_focus":153,"title_focus_zh":154,"keyword_extras":155,"keyword_extras_zh":158,"model_data":160},"claude-haiku-5-5",[5],[8,9],"openai","anthropic","anthropic\u002Fclaude-yunshang-awsb",{"input_tokens_tiers":12},{"tiers":13},[14,21],{"max_tokens":15,"pricing":16},100000,{"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":18,"output_tokens":24,"cache_read_tokens":25,"cache_write_tokens":26},"2.5","0.05","0.625",{"groups":28,"highlights":89,"overall":93,"scenario":95,"summary":98,"weakness":99},[29,68,81],{"items":30,"name":67},[31,37,42,46,51,55,59,63],{"detail":32,"method":33,"name":34,"samples":35,"score":36},"","Mixed","Agent Collaboration",18,81,{"detail":32,"method":38,"name":39,"samples":40,"score":41},"Objective","Code Generation",12,58,{"detail":32,"method":33,"name":43,"samples":44,"score":45},"Code Comprehension & Debugging",6,79,{"detail":32,"method":47,"name":48,"samples":49,"score":50},"Subjective","Creative Writing",22,89,{"detail":32,"method":33,"name":52,"samples":53,"score":54},"Information Retrieval \u002F RAG",9,92,{"detail":32,"method":47,"name":56,"samples":57,"score":58},"Long-Context Handling",16,83,{"detail":32,"method":33,"name":60,"samples":61,"score":62},"Math & Science Reasoning",14,78,{"detail":32,"method":47,"name":64,"samples":65,"score":66},"Multilingual Capability",11,91,"Advanced Capabilities",{"items":69,"name":80},[70,73,77],{"detail":32,"method":33,"name":71,"samples":72,"score":66},"General Knowledge",10,{"detail":32,"method":47,"name":74,"samples":75,"score":76},"Text Comprehension",23,82,{"detail":32,"method":33,"name":78,"samples":57,"score":79},"Logical Reasoning",80,"General Foundations",{"items":82,"name":88},[83,86],{"detail":32,"method":47,"name":84,"samples":85,"score":66},"Hallucination Control",15,{"detail":32,"method":47,"name":87,"samples":35,"score":76},"Instruction Following","Safety & Compliance",[90,91,92],{"name":52,"score":54},{"name":71,"score":66},{"name":64,"score":66},{"rating":94,"score":58},"良好",{"fit":96,"name":97},87,"E-commerce Copywriting \u002F Enterprise Knowledge Q&A \u002F Translation & Localization","Claude Haiku 5.5 scored 83 overall, performing well — a standout in the fast tier. It stands out in information retrieval \u002F RAG (92%), general knowledge (91%) and multilingual capability (91%), delivering retrieval-grounded, accurate answers at small-model speed and cost, which suits e-commerce copy, knowledge-base Q&A and localization at volume. Comparatively, code generation (58%) is weak; coding should go to bigger siblings. Big-model quality at small-model speed is its pitch.",{"name":39,"note":32,"score":41},{"groups":101,"highlights":140,"overall":144,"scenario":145,"summary":147,"weakness":148},[102,124,133],{"items":103,"name":123},[104,107,110,112,115,117,119,121],{"detail":32,"method":105,"name":106,"samples":35,"score":36},"主\u002F客观混合","智能体协同",{"detail":32,"method":108,"name":109,"samples":40,"score":41},"客观","代码生成",{"detail":32,"method":105,"name":111,"samples":44,"score":45},"代码理解与调试",{"detail":32,"method":113,"name":114,"samples":49,"score":50},"主观","创意写作",{"detail":32,"method":105,"name":116,"samples":53,"score":54},"信息检索 \u002F RAG",{"detail":32,"method":113,"name":118,"samples":57,"score":58},"长文本处理",{"detail":32,"method":105,"name":120,"samples":61,"score":62},"数学与科学推理",{"detail":32,"method":113,"name":122,"samples":65,"score":66},"多语言能力","专项进阶能力",{"items":125,"name":132},[126,128,130],{"detail":32,"method":105,"name":127,"samples":72,"score":66},"通识知识",{"detail":32,"method":113,"name":129,"samples":75,"score":76},"文本理解",{"detail":32,"method":105,"name":131,"samples":57,"score":79},"逻辑推理","通用基础能力",{"items":134,"name":139},[135,137],{"detail":32,"method":113,"name":136,"samples":85,"score":66},"幻觉控制",{"detail":32,"method":113,"name":138,"samples":35,"score":76},"指令遵循","安全与合规能力",[141,142,143],{"name":116,"score":54},{"name":127,"score":66},{"name":122,"score":66},{"rating":94,"score":58},{"fit":96,"name":146},"电商营销文案 \u002F 企业知识库问答 \u002F 翻译与本地化","Claude Haiku 5.5 综合得分 83，整体表现良好，是快速档中的佼佼者。信息检索 \u002F RAG（92%）、通识知识（91%）与多语言能力（91%）表现突出，以小模型的速度和成本输出检索有据的准确回答，在电商营销文案与知识库问答场景中优势明显。相对而言，代码生成（58%）偏弱，编码任务建议交给更大的同门模型。小模型速度、大模型质量是它的定位。",{"name":109,"note":32,"score":41},"Claude Haiku 5.5 is Anthropic's fast-tier model, pairing retrieval-grounded accuracy with broad knowledge and multilingual strength — built for high-volume Q&A, copywriting and localization.","Claude Haiku 5.5是Anthropic快速档模型，检索问答准确、知识广博、多语言能力强，适合高频问答、文案生成和本地化等规模化任务。","Claude Haiku 5.5 delivers RAG-grounded, multilingual answers at small-model speed and cost. Integrate in one click via the SilvaMux unified API.","Claude Haiku 5.5以小模型的速度和成本提供RAG级准确的多语言回答。可通过 SilvaMux 统一 API 一键快速接入。","Fast Accurate Model","快速准确模型",[156,157],"fast response","RAG",[159,157],"快速响应",{"summary":161,"recentAvailability":174,"observedAt":32,"bars":175,"metrics":179},[162,166,170],{"label":163,"value":164,"unit":165},"modelsDetail.performance.latency","38.96","s",{"label":167,"value":168,"unit":169},"modelsDetail.performance.throughput","75.98K","tokens\u002Fmin",{"label":171,"value":172,"unit":173},"modelsDetail.performance.availability","99.4","%","99.4%",{"degraded":176,"down":177},[],[178],20,[180,183,186,190,194,197],{"label":181,"value":182},"modelsDetail.performance.completionRate","99%",{"label":184,"value":185},"modelsDetail.performance.cacheHitRate","84%",{"label":187,"value":188,"unit":189},"modelsDetail.performance.tokenUsage","2,200,270,054","tokens",{"label":191,"value":192,"unit":193},"modelsDetail.performance.avgRequests","32","modelsDetail.performance.requestsUnit",{"label":195,"value":196,"unit":169},"modelsDetail.performance.tpm","75,982",{"label":198,"value":199,"unit":200},"modelsDetail.performance.qpm","683","req\u002Fmin",{"id":5,"aliases":202,"types":204,"pricing_group":10,"pricing":205,"display_name":5,"tier":212,"model_capabilities":213,"model_capabilities_zh":241,"introduction":149,"introduction_zh":150,"description":151,"description_zh":152,"title_focus":153,"title_focus_zh":154,"keyword_extras":269,"keyword_extras_zh":270,"model_data":271},[203],"claude-haiku-5-5@2601014",[9],{"input_tokens_tiers":206},{"tiers":207},[208,210],{"max_tokens":15,"pricing":209},{"input_tokens":17,"output_tokens":18,"cache_read_tokens":19,"cache_write_tokens":20},{"max_tokens":22,"pricing":211},{"input_tokens":18,"output_tokens":24,"cache_read_tokens":25,"cache_write_tokens":26},"2601014",{"groups":214,"highlights":234,"overall":238,"scenario":239,"summary":98,"weakness":240},[215,225,230],{"items":216,"name":67},[217,218,219,220,221,222,223,224],{"detail":32,"method":33,"name":34,"samples":35,"score":36},{"detail":32,"method":38,"name":39,"samples":40,"score":41},{"detail":32,"method":33,"name":43,"samples":44,"score":45},{"detail":32,"method":47,"name":48,"samples":49,"score":50},{"detail":32,"method":33,"name":52,"samples":53,"score":54},{"detail":32,"method":47,"name":56,"samples":57,"score":58},{"detail":32,"method":33,"name":60,"samples":61,"score":62},{"detail":32,"method":47,"name":64,"samples":65,"score":66},{"items":226,"name":80},[227,228,229],{"detail":32,"method":33,"name":71,"samples":72,"score":66},{"detail":32,"method":47,"name":74,"samples":75,"score":76},{"detail":32,"method":33,"name":78,"samples":57,"score":79},{"items":231,"name":88},[232,233],{"detail":32,"method":47,"name":84,"samples":85,"score":66},{"detail":32,"method":47,"name":87,"samples":35,"score":76},[235,236,237],{"name":52,"score":54},{"name":71,"score":66},{"name":64,"score":66},{"rating":94,"score":58},{"fit":96,"name":97},{"name":39,"note":32,"score":41},{"groups":242,"highlights":262,"overall":266,"scenario":267,"summary":147,"weakness":268},[243,253,258],{"items":244,"name":123},[245,246,247,248,249,250,251,252],{"detail":32,"method":105,"name":106,"samples":35,"score":36},{"detail":32,"method":108,"name":109,"samples":40,"score":41},{"detail":32,"method":105,"name":111,"samples":44,"score":45},{"detail":32,"method":113,"name":114,"samples":49,"score":50},{"detail":32,"method":105,"name":116,"samples":53,"score":54},{"detail":32,"method":113,"name":118,"samples":57,"score":58},{"detail":32,"method":105,"name":120,"samples":61,"score":62},{"detail":32,"method":113,"name":122,"samples":65,"score":66},{"items":254,"name":132},[255,256,257],{"detail":32,"method":105,"name":127,"samples":72,"score":66},{"detail":32,"method":113,"name":129,"samples":75,"score":76},{"detail":32,"method":105,"name":131,"samples":57,"score":79},{"items":259,"name":139},[260,261],{"detail":32,"method":113,"name":136,"samples":85,"score":66},{"detail":32,"method":113,"name":138,"samples":35,"score":76},[263,264,265],{"name":116,"score":54},{"name":127,"score":66},{"name":122,"score":66},{"rating":94,"score":58},{"fit":96,"name":146},{"name":109,"note":32,"score":41},[156,157],[159,157],{"summary":272,"recentAvailability":279,"observedAt":32,"bars":280,"metrics":286},[273,275,277],{"label":163,"value":274,"unit":165},"32.33",{"label":167,"value":276,"unit":169},"70.36K",{"label":171,"value":278,"unit":173},"99.2","99.2%",{"degraded":281,"down":283},[282],21,[284,285],4,24,[287,289,291,293,295,297],{"label":181,"value":288},"100%",{"label":184,"value":290},"88%",{"label":187,"value":292,"unit":189},"1,335,859,436",{"label":191,"value":294,"unit":193},"31",{"label":195,"value":296,"unit":169},"70,363",{"label":198,"value":298,"unit":200},"940",{"ttft":300,"tps":303,"e2e":306,"requests":309,"successfulRequests":310,"successRate":311,"availability":312},{"p50":301,"p95":302},1979,2312,{"p50":304,"p95":305},67,94,{"p50":307,"p95":308},6984,9286,67711,67364,"99.49",[313,314,313,313,313,313,313,313,313,313,313,313,313,313,313,314,314,313,313,313,313,313,313,313],"up","degraded"]