[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-kimi-k2-0905":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":143,"fitting_gpus":367,"community":368},13,"kimi-k2-0905","Kimi-K2-0905","moonshotai\u002FKimi-K2-0905","moonshotai","llama",null,1000000000000,"1000B",262144,"MODEL_LICENSE",280000,2400,"published","1T-parameter MoE agent model by Moonshot AI with 256K context.","https:\u002F\u002Fhuggingface.co\u002Fmoonshotai\u002FKimi-K2-0905","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},"Kimi","kimi",0.027,[],[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":83,"INT8":97,"AWQ":101,"GPTQ":115,"GGUF":129},{"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",4096,1862.65,2361.2,3541.79,2944.73,"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",16384,64,2688.05,4032.08,3009.73,{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":63,"kv_cache_gb":44,"vram_gb":80,"ram_gb":81,"disk_gb":82,"requirement_source":67,"confidence":68},"recommended",8192,2,2484.05,3726.07,2970.73,{"minimum":84,"production":89,"recommended":93},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":85,"kv_cache_gb":45,"vram_gb":86,"ram_gb":87,"disk_gb":88,"requirement_source":67,"confidence":68},931.32,1183.07,1774.61,1491.86,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":85,"kv_cache_gb":72,"vram_gb":90,"ram_gb":91,"disk_gb":92,"requirement_source":67,"confidence":68},1402.83,2104.24,1556.86,{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":85,"kv_cache_gb":44,"vram_gb":94,"ram_gb":95,"disk_gb":96,"requirement_source":67,"confidence":68},1252.37,1878.56,1517.86,{"minimum":98,"production":99,"recommended":100},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":85,"kv_cache_gb":45,"vram_gb":86,"ram_gb":87,"disk_gb":88,"requirement_source":67,"confidence":68},{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":85,"kv_cache_gb":72,"vram_gb":90,"ram_gb":91,"disk_gb":92,"requirement_source":67,"confidence":68},{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":85,"kv_cache_gb":44,"vram_gb":94,"ram_gb":95,"disk_gb":96,"requirement_source":67,"confidence":68},{"minimum":102,"production":107,"recommended":111},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":103,"kv_cache_gb":45,"vram_gb":104,"ram_gb":105,"disk_gb":106,"requirement_source":67,"confidence":68},483.12,616.1,924.15,792.67,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":103,"kv_cache_gb":72,"vram_gb":108,"ram_gb":109,"disk_gb":110,"requirement_source":67,"confidence":68},784.31,1176.47,857.67,{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":103,"kv_cache_gb":44,"vram_gb":112,"ram_gb":113,"disk_gb":114,"requirement_source":67,"confidence":68},659.63,989.45,818.67,{"minimum":116,"production":121,"recommended":125},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":117,"kv_cache_gb":45,"vram_gb":118,"ram_gb":119,"disk_gb":120,"requirement_source":67,"confidence":68},488.94,623.46,935.2,801.75,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":117,"kv_cache_gb":72,"vram_gb":122,"ram_gb":123,"disk_gb":124,"requirement_source":67,"confidence":68},792.34,1188.51,866.75,{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":117,"kv_cache_gb":44,"vram_gb":126,"ram_gb":127,"disk_gb":128,"requirement_source":67,"confidence":68},667.33,1000.99,827.75,{"minimum":130,"production":135,"recommended":139},{"scenario":61,"context_length":62,"batch_size":40,"weight_gb":131,"kv_cache_gb":45,"vram_gb":132,"ram_gb":133,"disk_gb":134,"requirement_source":67,"confidence":68},500.59,638.19,957.29,819.91,{"scenario":70,"context_length":71,"batch_size":34,"weight_gb":131,"kv_cache_gb":72,"vram_gb":136,"ram_gb":137,"disk_gb":138,"requirement_source":67,"confidence":68},808.41,1212.61,884.91,{"scenario":77,"context_length":78,"batch_size":79,"weight_gb":131,"kv_cache_gb":44,"vram_gb":140,"ram_gb":141,"disk_gb":142,"requirement_source":67,"confidence":68},682.72,1024.09,845.91,{"recommended":144,"cheapest":164,"performance":176,"min_complexity":179,"cluster_options":182,"no_match":365,"required_vram_gb":112,"candidate_count":79,"usage":366},{"gpu_model":145,"gpu_model_id":146,"gpu_count":44,"total_vram_gb":147,"hour_price":148,"monthly_cost":149,"providers":150,"bandwidth_gbps":152,"fp16_tflops":153,"interconnect_score":154,"parallel_efficiency":155,"score":156,"tier":157,"perf_metric":158,"perf_label":159,"perf_value":160,"workload":10,"cost_metric":161},"H200 SXM",11,1128,3.5,20440,[151],"lambda",4800,989,40,0.4,0.3757,"balanced","tok\u002Fs","31.8 tok\u002Fs",31.8,{"effective_value":162,"effective_unit":158,"capacity_factor":163},22.3,0.7,{"gpu_model":165,"gpu_model_id":166,"gpu_count":34,"total_vram_gb":167,"hour_price":168,"monthly_cost":169,"providers":170,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":172,"score":173,"tier":157,"perf_metric":158,"perf_label":174,"perf_value":41,"workload":10,"cost_metric":175},"B200",27,720,4.1271,12051.13,[171],"vast",0.6,0.3057,"0 tok\u002Fs",{"effective_value":41,"effective_unit":158,"capacity_factor":163},{"gpu_model":145,"gpu_model_id":146,"gpu_count":44,"total_vram_gb":147,"hour_price":148,"monthly_cost":149,"providers":177,"bandwidth_gbps":152,"fp16_tflops":153,"interconnect_score":154,"parallel_efficiency":155,"score":156,"tier":157,"perf_metric":158,"perf_label":159,"perf_value":160,"workload":10,"cost_metric":178},[151],{"effective_value":162,"effective_unit":158,"capacity_factor":163},{"gpu_model":165,"gpu_model_id":166,"gpu_count":34,"total_vram_gb":167,"hour_price":168,"monthly_cost":169,"providers":180,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":172,"score":173,"tier":157,"perf_metric":158,"perf_label":174,"perf_value":41,"workload":10,"cost_metric":181},[171],{"effective_value":41,"effective_unit":158,"capacity_factor":163},[183,197,206,217,226,233,240,250,261,269,277,285,294,301,311,316,325,332,341,346,357],{"gpu_model":184,"gpu_model_id":185,"gpu_count":44,"total_vram_gb":186,"hour_price":187,"monthly_cost":188,"providers":189,"bandwidth_gbps":192,"fp16_tflops":193,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":195,"score":196,"tier":157},"A100 80GB",9,640,1.2,7008,[171,190,191,151],"tensordock","runpod",2039,77.97,true,19.63,0.4073,{"gpu_model":198,"gpu_model_id":199,"gpu_count":44,"total_vram_gb":186,"hour_price":200,"monthly_cost":201,"providers":202,"bandwidth_gbps":203,"fp16_tflops":204,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":195,"score":205,"tier":157},"H100 SXM",10,1.6022,9356.85,[171,190,151,191],3350,66.9,0.3883,{"gpu_model":207,"gpu_model_id":208,"gpu_count":44,"total_vram_gb":209,"hour_price":210,"monthly_cost":211,"providers":212,"bandwidth_gbps":213,"fp16_tflops":214,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":215,"score":216,"tier":157},"RTX A6000",6,384,0.3001,1752.58,[171,151],768,77,275.63,0.4877,{"gpu_model":218,"gpu_model_id":219,"gpu_count":44,"total_vram_gb":209,"hour_price":220,"monthly_cost":221,"providers":222,"bandwidth_gbps":223,"fp16_tflops":224,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":215,"score":225,"tier":157},"L40S",12,0.4669,2726.7,[171,190,191],864,362,0.4696,{"gpu_model":227,"gpu_model_id":228,"gpu_count":44,"total_vram_gb":209,"hour_price":229,"monthly_cost":230,"providers":231,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":215,"score":232,"tier":157},"RTX PRO 6000 WS",19,0.8022,4684.85,[171],0.4091,{"gpu_model":234,"gpu_model_id":235,"gpu_count":44,"total_vram_gb":209,"hour_price":236,"monthly_cost":237,"providers":238,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":215,"score":239,"tier":157},"RTX PRO 6000 S",20,1.0131,5916.5,[171],0.3834,{"gpu_model":241,"gpu_model_id":79,"gpu_count":44,"total_vram_gb":242,"hour_price":243,"monthly_cost":244,"providers":245,"bandwidth_gbps":246,"fp16_tflops":247,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":248,"score":249,"tier":157},"RTX 5090",256,0.3222,1881.65,[171,191],1792,209,403.63,0.5084,{"gpu_model":251,"gpu_model_id":252,"gpu_count":44,"total_vram_gb":253,"hour_price":254,"monthly_cost":255,"providers":256,"bandwidth_gbps":257,"fp16_tflops":258,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":260,"tier":157},"RTX 3090",3,192,0.0678,395.95,[171,190],936,71,467.63,0.5198,{"gpu_model":262,"gpu_model_id":40,"gpu_count":44,"total_vram_gb":253,"hour_price":263,"monthly_cost":264,"providers":265,"bandwidth_gbps":266,"fp16_tflops":267,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":268,"tier":157},"RTX 4090",0.1344,784.9,[171,190,191,151],1008,165,0.5134,{"gpu_model":270,"gpu_model_id":271,"gpu_count":44,"total_vram_gb":253,"hour_price":272,"monthly_cost":273,"providers":274,"bandwidth_gbps":266,"fp16_tflops":275,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":276,"tier":157},"RTX 4090D",5,0.1602,935.57,[171],82.6,0.5102,{"gpu_model":278,"gpu_model_id":4,"gpu_count":44,"total_vram_gb":253,"hour_price":279,"monthly_cost":280,"providers":281,"bandwidth_gbps":282,"fp16_tflops":283,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":284,"tier":157},"L4",0.2015,1176.76,[171],300,121,0.489,{"gpu_model":286,"gpu_model_id":287,"gpu_count":44,"total_vram_gb":253,"hour_price":288,"monthly_cost":289,"providers":290,"bandwidth_gbps":291,"fp16_tflops":292,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":293,"tier":157},"A10",7,0.2409,1406.86,[171,191],600,125,0.4911,{"gpu_model":295,"gpu_model_id":296,"gpu_count":44,"total_vram_gb":253,"hour_price":297,"monthly_cost":298,"providers":299,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":259,"score":300,"tier":157},"RTX PRO 5000",26,0.3348,1955.23,[171],0.4659,{"gpu_model":302,"gpu_model_id":303,"gpu_count":44,"total_vram_gb":304,"hour_price":305,"monthly_cost":306,"providers":307,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":309,"tier":310},"Tesla V100",17,128,0.0272,158.85,[171],531.63,0.5034,"cheapest",{"gpu_model":312,"gpu_model_id":313,"gpu_count":44,"total_vram_gb":304,"hour_price":254,"monthly_cost":255,"providers":314,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":315,"tier":157},"RTX 4070S Ti",22,[171],0.4984,{"gpu_model":317,"gpu_model_id":318,"gpu_count":44,"total_vram_gb":304,"hour_price":319,"monthly_cost":320,"providers":321,"bandwidth_gbps":322,"fp16_tflops":323,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":324,"tier":157},"RTX 5080",23,0.1078,629.55,[171],960,113,0.5155,{"gpu_model":326,"gpu_model_id":327,"gpu_count":44,"total_vram_gb":304,"hour_price":328,"monthly_cost":329,"providers":330,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":331,"tier":157},"RTX 4080S",21,0.1209,706.06,[171],0.492,{"gpu_model":333,"gpu_model_id":334,"gpu_count":44,"total_vram_gb":304,"hour_price":335,"monthly_cost":336,"providers":337,"bandwidth_gbps":338,"fp16_tflops":339,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":340,"tier":157},"RTX 5070 Ti",18,0.1215,709.56,[171],896,88,0.5124,{"gpu_model":342,"gpu_model_id":343,"gpu_count":44,"total_vram_gb":304,"hour_price":243,"monthly_cost":244,"providers":344,"bandwidth_gbps":41,"fp16_tflops":41,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":308,"score":345,"tier":157},"RTX PRO 4500",25,[171],0.4675,{"gpu_model":347,"gpu_model_id":348,"gpu_count":44,"total_vram_gb":349,"hour_price":350,"monthly_cost":351,"providers":352,"bandwidth_gbps":353,"fp16_tflops":354,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":355,"score":356,"tier":157},"RTX 5070",28,96,0.0804,469.54,[171],672,61,563.63,0.5122,{"gpu_model":358,"gpu_model_id":359,"gpu_count":44,"total_vram_gb":72,"hour_price":360,"monthly_cost":361,"providers":362,"bandwidth_gbps":363,"fp16_tflops":359,"interconnect_score":154,"parallel_efficiency":155,"insufficient":194,"vram_shortfall_gb":364,"score":276,"tier":157},"RTX 5060 Ti",24,0.0554,323.54,[171],448,595.63,false,"value",[],{"posts_count":41,"benchmarks_count":41}]