[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-nvidia-nemotron-3-super":3},{"id":4,"slug":5,"name":6,"hf_id":7,"organization":8,"architecture":9,"architecture_config":10,"parameters":11,"parameters_label":12,"context_length":13,"license":14,"downloads":15,"likes":16,"status":17,"description":10,"huggingface_url":18,"created_at":19,"seo":20,"category":23,"family":10,"score":26,"variants":27,"quantizations":28,"requirements":55,"gpu_recommendations":140,"fitting_gpus":217,"community":233},262,"nvidia-nemotron-3-super","NVIDIA Nemotron 3 Super","nvidia\u002FNVIDIA-Nemotron-3-Super-120B-A12B-BF16","nvidia","moe_transformer",null,120000000000,"120B",1048576,"MIT",818925,415,"published","https:\u002F\u002Fhuggingface.co\u002Fnvidia\u002FNVIDIA-Nemotron-3-Super-120B-A12B-BF16","2026-08-12T03:11:19+00:00",{"title":6,"description":21,"image":10,"robots":22,"canonical_url":10},"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":24,"slug":25},"Coding AI","coding",0.1061,[],[29,34,39,44,48,52],{"format":30,"bits":31,"size_factor":32,"quality_loss":33},"AWQ",4,0.263,0.03,{"format":35,"bits":36,"size_factor":37,"quality_loss":38},"FP16",16,1,0,{"format":40,"bits":41,"size_factor":42,"quality_loss":43},"FP8",8,0.5,0.01,{"format":45,"bits":31,"size_factor":46,"quality_loss":47},"GGUF",0.25,0.05,{"format":49,"bits":31,"size_factor":50,"quality_loss":51},"GPTQ",0.275,0.04,{"format":53,"bits":41,"size_factor":42,"quality_loss":54},"INT8",0.02,{"FP16":56,"FP8":80,"INT8":94,"AWQ":98,"GPTQ":112,"GGUF":126},{"minimum":57,"production":66,"recommended":73},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":60,"kv_cache_gb":42,"vram_gb":61,"ram_gb":62,"disk_gb":63,"requirement_source":64,"confidence":65},"minimum",4096,223.52,287.7,431.55,387.69,"estimated",0.95,{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":60,"kv_cache_gb":69,"vram_gb":70,"ram_gb":71,"disk_gb":72,"requirement_source":64,"confidence":65},"production",16384,64,426.05,639.08,452.69,{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":60,"kv_cache_gb":41,"vram_gb":77,"ram_gb":78,"disk_gb":79,"requirement_source":64,"confidence":65},"recommended",8192,2,316.3,474.45,413.69,{"minimum":81,"production":86,"recommended":90},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":82,"kv_cache_gb":42,"vram_gb":83,"ram_gb":84,"disk_gb":85,"requirement_source":64,"confidence":65},111.76,146.32,219.49,213.34,{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":82,"kv_cache_gb":69,"vram_gb":87,"ram_gb":88,"disk_gb":89,"requirement_source":64,"confidence":65},271.83,407.74,278.34,{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":82,"kv_cache_gb":41,"vram_gb":91,"ram_gb":92,"disk_gb":93,"requirement_source":64,"confidence":65},168.5,252.75,239.34,{"minimum":95,"production":96,"recommended":97},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":82,"kv_cache_gb":42,"vram_gb":83,"ram_gb":84,"disk_gb":85,"requirement_source":64,"confidence":65},{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":82,"kv_cache_gb":69,"vram_gb":87,"ram_gb":88,"disk_gb":89,"requirement_source":64,"confidence":65},{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":82,"kv_cache_gb":41,"vram_gb":91,"ram_gb":92,"disk_gb":93,"requirement_source":64,"confidence":65},{"minimum":99,"production":104,"recommended":108},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":100,"kv_cache_gb":42,"vram_gb":101,"ram_gb":102,"disk_gb":103,"requirement_source":64,"confidence":65},57.97,78.29,117.43,129.44,{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":100,"kv_cache_gb":69,"vram_gb":105,"ram_gb":106,"disk_gb":107,"requirement_source":64,"confidence":65},197.61,296.41,194.44,{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":100,"kv_cache_gb":41,"vram_gb":109,"ram_gb":110,"disk_gb":111,"requirement_source":64,"confidence":65},97.37,146.06,155.44,{"minimum":113,"production":118,"recommended":122},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":114,"kv_cache_gb":42,"vram_gb":115,"ram_gb":116,"disk_gb":117,"requirement_source":64,"confidence":65},58.67,79.17,118.76,130.53,{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":114,"kv_cache_gb":69,"vram_gb":119,"ram_gb":120,"disk_gb":121,"requirement_source":64,"confidence":65},198.57,297.85,195.53,{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":114,"kv_cache_gb":41,"vram_gb":123,"ram_gb":124,"disk_gb":125,"requirement_source":64,"confidence":65},98.3,147.44,156.53,{"minimum":127,"production":132,"recommended":136},{"scenario":58,"context_length":59,"batch_size":37,"weight_gb":128,"kv_cache_gb":42,"vram_gb":129,"ram_gb":130,"disk_gb":131,"requirement_source":64,"confidence":65},60.07,80.94,121.41,132.71,{"scenario":67,"context_length":68,"batch_size":31,"weight_gb":128,"kv_cache_gb":69,"vram_gb":133,"ram_gb":134,"disk_gb":135,"requirement_source":64,"confidence":65},200.5,300.75,197.71,{"scenario":74,"context_length":75,"batch_size":76,"weight_gb":128,"kv_cache_gb":41,"vram_gb":137,"ram_gb":138,"disk_gb":139,"requirement_source":64,"confidence":65},100.14,150.21,158.71,{"recommended":141,"cheapest":161,"performance":174,"min_complexity":183,"cluster_options":192,"no_match":214,"required_vram_gb":109,"candidate_count":215,"usage":216},{"gpu_model":142,"gpu_model_id":143,"gpu_count":31,"total_vram_gb":144,"hour_price":145,"monthly_cost":146,"providers":147,"bandwidth_gbps":149,"fp16_tflops":150,"interconnect_score":151,"parallel_efficiency":152,"score":153,"tier":154,"perf_metric":155,"perf_label":156,"perf_value":157,"workload":10,"cost_metric":158},"H200 SXM",11,564,3.5,10220,[148],"lambda",4800,989,40,0.6,0.3154,"balanced","tok\u002Fs","198.7 tok\u002Fs",198.7,{"effective_value":159,"effective_unit":155,"capacity_factor":160},139.1,0.7,{"gpu_model":162,"gpu_model_id":163,"gpu_count":41,"total_vram_gb":164,"hour_price":165,"monthly_cost":166,"providers":167,"bandwidth_gbps":38,"fp16_tflops":38,"interconnect_score":151,"parallel_efficiency":169,"score":170,"tier":171,"perf_metric":155,"perf_label":172,"perf_value":38,"workload":10,"cost_metric":173},"Tesla V100",17,128,0.0272,158.85,[168],"vast",0.4,0.5538,"cheapest","0 tok\u002Fs",{"effective_value":38,"effective_unit":155,"capacity_factor":160},{"gpu_model":142,"gpu_model_id":143,"gpu_count":41,"total_vram_gb":175,"hour_price":145,"monthly_cost":176,"providers":177,"bandwidth_gbps":149,"fp16_tflops":150,"interconnect_score":151,"parallel_efficiency":169,"score":178,"tier":154,"perf_metric":155,"perf_label":179,"perf_value":180,"workload":10,"cost_metric":181},1128,20440,[148],0.2197,"265 tok\u002Fs",265,{"effective_value":182,"effective_unit":155,"capacity_factor":160},185.5,{"gpu_model":142,"gpu_model_id":143,"gpu_count":37,"total_vram_gb":184,"hour_price":145,"monthly_cost":185,"providers":186,"bandwidth_gbps":149,"fp16_tflops":150,"interconnect_score":151,"parallel_efficiency":37,"score":187,"tier":154,"perf_metric":155,"perf_label":188,"perf_value":189,"workload":10,"cost_metric":190},141,2555,[148],0.6552,"82.8 tok\u002Fs",82.8,{"effective_value":191,"effective_unit":155,"capacity_factor":160},58,[193,205],{"gpu_model":194,"gpu_model_id":195,"gpu_count":41,"total_vram_gb":196,"hour_price":197,"monthly_cost":198,"providers":199,"bandwidth_gbps":200,"fp16_tflops":201,"interconnect_score":151,"parallel_efficiency":169,"insufficient":202,"vram_shortfall_gb":203,"score":204,"tier":171},"RTX 5070",28,96,0.0101,58.98,[168],672,61,true,1.37,0.5208,{"gpu_model":206,"gpu_model_id":207,"gpu_count":41,"total_vram_gb":69,"hour_price":208,"monthly_cost":209,"providers":210,"bandwidth_gbps":211,"fp16_tflops":207,"interconnect_score":151,"parallel_efficiency":169,"insufficient":202,"vram_shortfall_gb":212,"score":213,"tier":154},"RTX 5060 Ti",24,0.0544,317.7,[168],448,33.37,0.5103,false,14,"value",[218,222,230],{"name":142,"vram_gb":184,"cheapest_hour":145,"cost":219,"provider":148,"estimated_tps":189},{"hourly":145,"daily":220,"monthly":185,"yearly":221},84,30660,{"name":223,"vram_gb":224,"cheapest_hour":225,"cost":226,"provider":168,"estimated_tps":38},"B200",180,4.1271,{"hourly":225,"daily":227,"monthly":228,"yearly":229},99.05,3012.78,36153.4,{"name":231,"vram_gb":224,"cheapest_hour":10,"cost":10,"provider":10,"estimated_tps":232},"B200 SXM",134.6,{"posts_count":38,"benchmarks_count":38}]