[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-dolphin-2-9-1-yi-1-5-34b":3},{"id":4,"slug":5,"name":6,"hf_id":6,"organization":7,"architecture":8,"architecture_config":9,"parameters":14,"parameters_label":15,"context_length":16,"license":17,"downloads":18,"likes":19,"status":20,"description":17,"huggingface_url":21,"created_at":22,"seo":23,"category":26,"family":29,"score":31,"variants":32,"quantizations":33,"requirements":59,"gpu_recommendations":145,"fitting_gpus":217,"community":245},501,"dolphin-2-9-1-yi-1-5-34b","dphn\u002Fdolphin-2.9.1-yi-1.5-34b","dphn","llama",{"layers":10,"head_dim":11,"kv_heads":12,"hidden_size":13},60,128,8,7168,37452513280,"37.5B",8192,null,4527536,65,"published","https:\u002F\u002Fhuggingface.co\u002Fdphn\u002Fdolphin-2.9.1-yi-1.5-34b","2026-08-13T18:12:18+00:00",{"title":6,"description":24,"image":17,"robots":25,"canonical_url":17},"AI 模型部署决策引擎——开源 AI 的 GPU 需求、云端价格与成本分析。","index, follow",{"name":27,"slug":28},"LLM","llm",{"name":30,"slug":8},"Llama",0.0078,[],[34,39,44,48,52,56],{"format":35,"bits":36,"size_factor":37,"quality_loss":38},"AWQ",4,0.263,0.03,{"format":40,"bits":41,"size_factor":42,"quality_loss":43},"FP16",16,1,0,{"format":45,"bits":12,"size_factor":46,"quality_loss":47},"FP8",0.5,0.01,{"format":49,"bits":36,"size_factor":50,"quality_loss":51},"GGUF",0.25,0.05,{"format":53,"bits":36,"size_factor":54,"quality_loss":55},"GPTQ",0.275,0.04,{"format":57,"bits":12,"size_factor":46,"quality_loss":58},"INT8",0.02,{"FP16":60,"FP8":85,"INT8":99,"AWQ":103,"GPTQ":117,"GGUF":131},{"minimum":61,"production":71,"recommended":78},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":64,"kv_cache_gb":65,"vram_gb":66,"ram_gb":67,"disk_gb":68,"requirement_source":69,"confidence":70},"minimum",4096,69.76,0.94,98.7,148.05,147.83,"inferred",0.95,{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":64,"kv_cache_gb":74,"vram_gb":75,"ram_gb":76,"disk_gb":77,"requirement_source":69,"confidence":70},"production",16384,120,368.67,553,212.83,{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":64,"kv_cache_gb":81,"vram_gb":82,"ram_gb":83,"disk_gb":84,"requirement_source":69,"confidence":70},"recommended",2,15,142,212.99,173.83,{"minimum":86,"production":91,"recommended":95},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":87,"kv_cache_gb":65,"vram_gb":88,"ram_gb":89,"disk_gb":90,"requirement_source":69,"confidence":70},34.88,54.57,81.86,93.41,{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":87,"kv_cache_gb":74,"vram_gb":92,"ram_gb":93,"disk_gb":94,"requirement_source":69,"confidence":70},320.53,480.8,158.41,{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":87,"kv_cache_gb":81,"vram_gb":96,"ram_gb":97,"disk_gb":98,"requirement_source":69,"confidence":70},95.87,143.8,119.41,{"minimum":100,"production":101,"recommended":102},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":87,"kv_cache_gb":65,"vram_gb":88,"ram_gb":89,"disk_gb":90,"requirement_source":69,"confidence":70},{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":87,"kv_cache_gb":74,"vram_gb":92,"ram_gb":93,"disk_gb":94,"requirement_source":69,"confidence":70},{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":87,"kv_cache_gb":81,"vram_gb":96,"ram_gb":97,"disk_gb":98,"requirement_source":69,"confidence":70},{"minimum":104,"production":109,"recommended":113},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":105,"kv_cache_gb":65,"vram_gb":106,"ram_gb":107,"disk_gb":108,"requirement_source":69,"confidence":70},18.09,33.34,50.01,67.23,{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":105,"kv_cache_gb":74,"vram_gb":110,"ram_gb":111,"disk_gb":112,"requirement_source":69,"confidence":70},297.37,446.05,132.23,{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":105,"kv_cache_gb":81,"vram_gb":114,"ram_gb":115,"disk_gb":116,"requirement_source":69,"confidence":70},73.67,110.5,93.23,{"minimum":118,"production":123,"recommended":127},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":119,"kv_cache_gb":65,"vram_gb":120,"ram_gb":121,"disk_gb":122,"requirement_source":69,"confidence":70},18.31,33.61,50.42,67.57,{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":119,"kv_cache_gb":74,"vram_gb":124,"ram_gb":125,"disk_gb":126,"requirement_source":69,"confidence":70},297.67,446.51,132.57,{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":119,"kv_cache_gb":81,"vram_gb":128,"ram_gb":129,"disk_gb":130,"requirement_source":69,"confidence":70},73.96,110.93,93.57,{"minimum":132,"production":137,"recommended":141},{"scenario":62,"context_length":63,"batch_size":42,"weight_gb":133,"kv_cache_gb":65,"vram_gb":134,"ram_gb":135,"disk_gb":136,"requirement_source":69,"confidence":70},18.75,34.17,51.25,68.25,{"scenario":72,"context_length":73,"batch_size":36,"weight_gb":133,"kv_cache_gb":74,"vram_gb":138,"ram_gb":139,"disk_gb":140,"requirement_source":69,"confidence":70},298.27,447.41,133.25,{"scenario":79,"context_length":16,"batch_size":80,"weight_gb":133,"kv_cache_gb":81,"vram_gb":142,"ram_gb":143,"disk_gb":144,"requirement_source":69,"confidence":70},74.53,111.8,94.25,{"recommended":146,"cheapest":165,"performance":177,"min_complexity":186,"cluster_options":202,"no_match":214,"required_vram_gb":114,"candidate_count":215,"usage":216},{"gpu_model":147,"gpu_model_id":148,"gpu_count":42,"total_vram_gb":149,"hour_price":150,"monthly_cost":151,"providers":152,"bandwidth_gbps":154,"fp16_tflops":155,"interconnect_score":156,"parallel_efficiency":42,"score":157,"tier":158,"perf_metric":159,"perf_label":160,"perf_value":161,"workload":17,"cost_metric":162},"H200 SXM",11,141,3.5,2555,[153],"lambda",4800,989,40,0.6104,"balanced","tok\u002Fs","265.3 tok\u002Fs",265.3,{"effective_value":163,"effective_unit":159,"capacity_factor":164},185.7,0.7,{"gpu_model":166,"gpu_model_id":167,"gpu_count":12,"total_vram_gb":11,"hour_price":168,"monthly_cost":169,"providers":170,"bandwidth_gbps":43,"fp16_tflops":43,"interconnect_score":156,"parallel_efficiency":172,"score":173,"tier":174,"perf_metric":159,"perf_label":175,"perf_value":43,"workload":17,"cost_metric":176},"Tesla V100",17,0.0289,168.78,[171],"vast",0.4,0.51,"cheapest","0 tok\u002Fs",{"effective_value":43,"effective_unit":159,"capacity_factor":164},{"gpu_model":147,"gpu_model_id":148,"gpu_count":12,"total_vram_gb":178,"hour_price":150,"monthly_cost":179,"providers":180,"bandwidth_gbps":154,"fp16_tflops":155,"interconnect_score":156,"parallel_efficiency":172,"score":181,"tier":158,"perf_metric":159,"perf_label":182,"perf_value":183,"workload":17,"cost_metric":184},1128,20440,[153],0.2197,"849.1 tok\u002Fs",849.1,{"effective_value":185,"effective_unit":159,"capacity_factor":164},594.4,{"gpu_model":187,"gpu_model_id":188,"gpu_count":42,"total_vram_gb":189,"hour_price":190,"monthly_cost":191,"providers":192,"bandwidth_gbps":195,"fp16_tflops":196,"interconnect_score":156,"parallel_efficiency":42,"score":197,"tier":158,"perf_metric":159,"perf_label":198,"perf_value":199,"workload":17,"cost_metric":200},"A100 80GB",9,80,1.2,876,[171,193,194,153],"tensordock","runpod",2039,77.97,0.6407,"112.7 tok\u002Fs",112.7,{"effective_value":201,"effective_unit":159,"capacity_factor":164},78.9,[203],{"gpu_model":204,"gpu_model_id":205,"gpu_count":12,"total_vram_gb":206,"hour_price":207,"monthly_cost":208,"providers":209,"bandwidth_gbps":210,"fp16_tflops":205,"interconnect_score":156,"parallel_efficiency":172,"insufficient":211,"vram_shortfall_gb":212,"score":213,"tier":158},"RTX 5060 Ti",24,64,0.0719,419.9,[171],448,true,9.67,0.5082,false,19,"value",[218,222,230,234,242],{"name":187,"vram_gb":189,"cheapest_hour":190,"cost":219,"provider":171,"estimated_tps":199},{"hourly":190,"daily":220,"monthly":191,"yearly":221},28.8,10512,{"name":223,"vram_gb":189,"cheapest_hour":224,"cost":225,"provider":171,"estimated_tps":229},"H100 SXM",1.6139,{"hourly":224,"daily":226,"monthly":227,"yearly":228},38.73,1178.15,14137.76,185.2,{"name":147,"vram_gb":149,"cheapest_hour":150,"cost":231,"provider":153,"estimated_tps":161},{"hourly":150,"daily":232,"monthly":151,"yearly":233},84,30660,{"name":235,"vram_gb":236,"cheapest_hour":237,"cost":238,"provider":171,"estimated_tps":43},"B200",180,4.1271,{"hourly":237,"daily":239,"monthly":240,"yearly":241},99.05,3012.78,36153.4,{"name":243,"vram_gb":236,"cheapest_hour":17,"cost":17,"provider":17,"estimated_tps":244},"B200 SXM",431.2,{"posts_count":43,"benchmarks_count":43}]