[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-deepseek-v3-1":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":18,"huggingface_url":19,"created_at":20,"seo":21,"category":23,"family":26,"score":29,"variants":30,"quantizations":31,"requirements":58,"gpu_recommendations":142,"fitting_gpus":357,"community":358},2,"deepseek-v3-1","DeepSeek V3","deepseek-ai\u002FDeepSeek-V3","deepseek-ai","moe_transformer",null,671000000000,"671B",131072,"MIT",7200000,9800,"published","General-purpose MoE LLM with 671B total parameters, strong coding and Chinese capabilities.","https:\u002F\u002Fhuggingface.co\u002Fdeepseek-ai\u002FDeepSeek-V3","2026-08-12T02:22:23+00:00",{"title":6,"description":18,"image":10,"robots":22,"canonical_url":10},"index, follow",{"name":24,"slug":25},"LLM","llm",{"name":27,"slug":28},"DeepSeek","deepseek",0.1615,[],[32,37,42,47,51,55],{"format":33,"bits":34,"size_factor":35,"quality_loss":36},"AWQ",4,0.263,0.03,{"format":38,"bits":39,"size_factor":40,"quality_loss":41},"FP16",16,1,0,{"format":43,"bits":44,"size_factor":45,"quality_loss":46},"FP8",8,0.5,0.01,{"format":48,"bits":34,"size_factor":49,"quality_loss":50},"GGUF",0.25,0.05,{"format":52,"bits":34,"size_factor":53,"quality_loss":54},"GPTQ",0.275,0.04,{"format":56,"bits":44,"size_factor":45,"quality_loss":57},"INT8",0.02,{"FP16":59,"FP8":82,"INT8":96,"AWQ":100,"GPTQ":114,"GGUF":128},{"minimum":60,"production":69,"recommended":76},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":63,"kv_cache_gb":45,"vram_gb":64,"ram_gb":65,"disk_gb":66,"requirement_source":67,"confidence":68},"minimum",4096,1249.83,1585.99,2378.99,1988.74,"estimated",0.95,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":63,"kv_cache_gb":72,"vram_gb":73,"ram_gb":74,"disk_gb":75,"requirement_source":67,"confidence":68},"production",16384,64,1842.37,2763.56,2053.74,{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":63,"kv_cache_gb":44,"vram_gb":79,"ram_gb":80,"disk_gb":81,"requirement_source":67,"confidence":68},"recommended",8192,1673.61,2510.41,2014.74,{"minimum":83,"production":88,"recommended":92},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":84,"kv_cache_gb":45,"vram_gb":85,"ram_gb":86,"disk_gb":87,"requirement_source":67,"confidence":68},624.92,795.47,1193.21,1013.87,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":84,"kv_cache_gb":72,"vram_gb":89,"ram_gb":90,"disk_gb":91,"requirement_source":67,"confidence":68},979.99,1469.98,1078.87,{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":84,"kv_cache_gb":44,"vram_gb":93,"ram_gb":94,"disk_gb":95,"requirement_source":67,"confidence":68},847.15,1270.73,1039.87,{"minimum":97,"production":98,"recommended":99},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":84,"kv_cache_gb":45,"vram_gb":85,"ram_gb":86,"disk_gb":87,"requirement_source":67,"confidence":68},{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":84,"kv_cache_gb":72,"vram_gb":89,"ram_gb":90,"disk_gb":91,"requirement_source":67,"confidence":68},{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":84,"kv_cache_gb":44,"vram_gb":93,"ram_gb":94,"disk_gb":95,"requirement_source":67,"confidence":68},{"minimum":101,"production":106,"recommended":110},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":102,"kv_cache_gb":45,"vram_gb":103,"ram_gb":104,"disk_gb":105,"requirement_source":67,"confidence":68},324.18,415.03,622.55,544.71,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":102,"kv_cache_gb":72,"vram_gb":107,"ram_gb":108,"disk_gb":109,"requirement_source":67,"confidence":68},564.96,847.44,609.71,{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":102,"kv_cache_gb":44,"vram_gb":111,"ram_gb":112,"disk_gb":113,"requirement_source":67,"confidence":68},449.42,674.13,570.71,{"minimum":115,"production":120,"recommended":124},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":116,"kv_cache_gb":45,"vram_gb":117,"ram_gb":118,"disk_gb":119,"requirement_source":67,"confidence":68},328.08,419.97,629.96,550.81,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":116,"kv_cache_gb":72,"vram_gb":121,"ram_gb":122,"disk_gb":123,"requirement_source":67,"confidence":68},570.35,855.53,615.81,{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":116,"kv_cache_gb":44,"vram_gb":125,"ram_gb":126,"disk_gb":127,"requirement_source":67,"confidence":68},454.59,681.88,576.81,{"minimum":129,"production":134,"recommended":138},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":130,"kv_cache_gb":45,"vram_gb":131,"ram_gb":132,"disk_gb":133,"requirement_source":67,"confidence":68},335.89,429.85,644.78,562.99,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":130,"kv_cache_gb":72,"vram_gb":135,"ram_gb":136,"disk_gb":137,"requirement_source":67,"confidence":68},581.13,871.7,627.99,{"scenario":77,"context_length":78,"batch_size":4,"weight_gb":130,"kv_cache_gb":44,"vram_gb":139,"ram_gb":140,"disk_gb":141,"requirement_source":67,"confidence":68},464.92,697.38,588.99,{"recommended":143,"cheapest":163,"performance":180,"min_complexity":190,"cluster_options":193,"no_match":355,"required_vram_gb":111,"candidate_count":34,"usage":356},{"gpu_model":144,"gpu_model_id":145,"gpu_count":44,"total_vram_gb":146,"hour_price":147,"monthly_cost":148,"providers":149,"bandwidth_gbps":151,"fp16_tflops":152,"interconnect_score":153,"parallel_efficiency":154,"score":155,"tier":156,"perf_metric":157,"perf_label":158,"perf_value":159,"workload":10,"cost_metric":160},"H200 SXM",11,1128,3.5,20440,[150],"lambda",4800,989,40,0.4,0.326,"balanced","tok\u002Fs","47.4 tok\u002Fs",47.4,{"effective_value":161,"effective_unit":157,"capacity_factor":162},33.2,0.7,{"gpu_model":164,"gpu_model_id":165,"gpu_count":44,"total_vram_gb":166,"hour_price":167,"monthly_cost":168,"providers":169,"bandwidth_gbps":173,"fp16_tflops":174,"interconnect_score":153,"parallel_efficiency":154,"score":175,"tier":156,"perf_metric":157,"perf_label":176,"perf_value":177,"workload":10,"cost_metric":178},"A100 80GB",9,640,1.2,7008,[170,171,172,150],"vast","tensordock","runpod",2039,77.97,0.4479,"20.1 tok\u002Fs",20.1,{"effective_value":179,"effective_unit":157,"capacity_factor":162},14.1,{"gpu_model":144,"gpu_model_id":145,"gpu_count":34,"total_vram_gb":181,"hour_price":147,"monthly_cost":182,"providers":183,"bandwidth_gbps":151,"fp16_tflops":152,"interconnect_score":153,"parallel_efficiency":184,"score":185,"tier":156,"perf_metric":157,"perf_label":186,"perf_value":187,"workload":10,"cost_metric":188},564,10220,[150],0.6,0.4569,"35.5 tok\u002Fs",35.5,{"effective_value":189,"effective_unit":157,"capacity_factor":162},24.8,{"gpu_model":144,"gpu_model_id":145,"gpu_count":34,"total_vram_gb":181,"hour_price":147,"monthly_cost":182,"providers":191,"bandwidth_gbps":151,"fp16_tflops":152,"interconnect_score":153,"parallel_efficiency":184,"score":185,"tier":156,"perf_metric":157,"perf_label":186,"perf_value":187,"workload":10,"cost_metric":192},[150],{"effective_value":189,"effective_unit":157,"capacity_factor":162},[194,206,215,222,229,239,250,258,266,275,284,291,301,306,315,322,331,336,347],{"gpu_model":195,"gpu_model_id":196,"gpu_count":44,"total_vram_gb":197,"hour_price":198,"monthly_cost":199,"providers":200,"bandwidth_gbps":201,"fp16_tflops":202,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":204,"score":205,"tier":156},"RTX A6000",6,384,0.3001,1752.58,[170,150],768,77,true,65.42,0.4877,{"gpu_model":207,"gpu_model_id":208,"gpu_count":44,"total_vram_gb":197,"hour_price":209,"monthly_cost":210,"providers":211,"bandwidth_gbps":212,"fp16_tflops":213,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":204,"score":214,"tier":156},"L40S",12,0.4669,2726.7,[170,171,172],864,362,0.4696,{"gpu_model":216,"gpu_model_id":217,"gpu_count":44,"total_vram_gb":197,"hour_price":218,"monthly_cost":219,"providers":220,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":204,"score":221,"tier":156},"RTX PRO 6000 WS",19,0.8022,4684.85,[170],0.4091,{"gpu_model":223,"gpu_model_id":224,"gpu_count":44,"total_vram_gb":197,"hour_price":225,"monthly_cost":226,"providers":227,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":204,"score":228,"tier":156},"RTX PRO 6000 S",20,1.0131,5916.5,[170],0.3834,{"gpu_model":230,"gpu_model_id":4,"gpu_count":44,"total_vram_gb":231,"hour_price":232,"monthly_cost":233,"providers":234,"bandwidth_gbps":235,"fp16_tflops":236,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":237,"score":238,"tier":156},"RTX 5090",256,0.3222,1881.65,[170,172],1792,209,193.42,0.5084,{"gpu_model":240,"gpu_model_id":241,"gpu_count":44,"total_vram_gb":242,"hour_price":243,"monthly_cost":244,"providers":245,"bandwidth_gbps":246,"fp16_tflops":247,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":249,"tier":156},"RTX 3090",3,192,0.0678,395.95,[170,171],936,71,257.42,0.5198,{"gpu_model":251,"gpu_model_id":40,"gpu_count":44,"total_vram_gb":242,"hour_price":252,"monthly_cost":253,"providers":254,"bandwidth_gbps":255,"fp16_tflops":256,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":257,"tier":156},"RTX 4090",0.1344,784.9,[170,171,172,150],1008,165,0.5134,{"gpu_model":259,"gpu_model_id":260,"gpu_count":44,"total_vram_gb":242,"hour_price":261,"monthly_cost":262,"providers":263,"bandwidth_gbps":255,"fp16_tflops":264,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":265,"tier":156},"RTX 4090D",5,0.1602,935.57,[170],82.6,0.5102,{"gpu_model":267,"gpu_model_id":268,"gpu_count":44,"total_vram_gb":242,"hour_price":269,"monthly_cost":270,"providers":271,"bandwidth_gbps":272,"fp16_tflops":273,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":274,"tier":156},"L4",13,0.2015,1176.76,[170],300,121,0.489,{"gpu_model":276,"gpu_model_id":277,"gpu_count":44,"total_vram_gb":242,"hour_price":278,"monthly_cost":279,"providers":280,"bandwidth_gbps":281,"fp16_tflops":282,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":283,"tier":156},"A10",7,0.2409,1406.86,[170,172],600,125,0.4911,{"gpu_model":285,"gpu_model_id":286,"gpu_count":44,"total_vram_gb":242,"hour_price":287,"monthly_cost":288,"providers":289,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":248,"score":290,"tier":156},"RTX PRO 5000",26,0.3348,1955.23,[170],0.4659,{"gpu_model":292,"gpu_model_id":293,"gpu_count":44,"total_vram_gb":294,"hour_price":295,"monthly_cost":296,"providers":297,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":299,"tier":300},"Tesla V100",17,128,0.0272,158.85,[170],321.42,0.5034,"cheapest",{"gpu_model":302,"gpu_model_id":303,"gpu_count":44,"total_vram_gb":294,"hour_price":243,"monthly_cost":244,"providers":304,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":305,"tier":156},"RTX 4070S Ti",22,[170],0.4984,{"gpu_model":307,"gpu_model_id":308,"gpu_count":44,"total_vram_gb":294,"hour_price":309,"monthly_cost":310,"providers":311,"bandwidth_gbps":312,"fp16_tflops":313,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":314,"tier":156},"RTX 5080",23,0.1078,629.55,[170],960,113,0.5155,{"gpu_model":316,"gpu_model_id":317,"gpu_count":44,"total_vram_gb":294,"hour_price":318,"monthly_cost":319,"providers":320,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":321,"tier":156},"RTX 4080S",21,0.1209,706.06,[170],0.492,{"gpu_model":323,"gpu_model_id":324,"gpu_count":44,"total_vram_gb":294,"hour_price":325,"monthly_cost":326,"providers":327,"bandwidth_gbps":328,"fp16_tflops":329,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":330,"tier":156},"RTX 5070 Ti",18,0.1215,709.56,[170],896,88,0.5124,{"gpu_model":332,"gpu_model_id":333,"gpu_count":44,"total_vram_gb":294,"hour_price":232,"monthly_cost":233,"providers":334,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":298,"score":335,"tier":156},"RTX PRO 4500",25,[170],0.4675,{"gpu_model":337,"gpu_model_id":338,"gpu_count":44,"total_vram_gb":339,"hour_price":340,"monthly_cost":341,"providers":342,"bandwidth_gbps":343,"fp16_tflops":344,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":345,"score":346,"tier":156},"RTX 5070",28,96,0.0804,469.54,[170],672,61,353.42,0.5122,{"gpu_model":348,"gpu_model_id":349,"gpu_count":44,"total_vram_gb":72,"hour_price":350,"monthly_cost":351,"providers":352,"bandwidth_gbps":353,"fp16_tflops":349,"interconnect_score":153,"parallel_efficiency":154,"insufficient":203,"vram_shortfall_gb":354,"score":265,"tier":156},"RTX 5060 Ti",24,0.0554,323.54,[170],448,385.42,false,"value",[],{"posts_count":41,"benchmarks_count":41}]