[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-granite-code-34b":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":211,"community":246},172,"granite-code-34b","Granite Code 34B","ibm-granite\u002Fgranite-34b-code-instruct","ibm-granite","transformer",null,34000000000,"34B",131072,"MIT",590,78,"published","https:\u002F\u002Fhuggingface.co\u002Fibm-granite\u002Fgranite-34b-code-instruct","2026-08-12T03:11:15+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",0.0009,[],[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",4096,63.33,85.06,127.59,137.79,"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",16384,64,205,307.49,202.79,{"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",8192,2,104.45,156.68,163.79,{"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},31.66,45.01,67.51,88.4,{"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},161.3,241.95,153.4,{"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},62.58,93.87,114.4,{"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},16.43,25.73,38.59,64.62,{"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},140.27,210.4,129.62,{"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},42.42,63.64,90.62,{"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},16.62,25.98,38.97,64.93,{"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},140.54,210.81,129.93,{"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},42.69,64.03,90.93,{"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},17.02,26.48,39.72,65.55,{"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},141.09,211.63,130.55,{"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},43.21,64.81,91.55,{"recommended":141,"cheapest":163,"performance":179,"min_complexity":194,"cluster_options":207,"no_match":208,"required_vram_gb":109,"candidate_count":209,"usage":210},{"gpu_model":142,"gpu_model_id":143,"gpu_count":37,"total_vram_gb":144,"hour_price":145,"monthly_cost":146,"providers":147,"bandwidth_gbps":152,"fp16_tflops":153,"interconnect_score":154,"parallel_efficiency":37,"score":155,"tier":156,"perf_metric":157,"perf_label":158,"perf_value":159,"workload":10,"cost_metric":160},"H100 SXM",10,80,1.6059,1172.31,[148,149,150,151],"vast","tensordock","runpod","lambda",3350,66.9,40,0.6309,"balanced","tok\u002Fs","203.9 tok\u002Fs",203.9,{"effective_value":161,"effective_unit":157,"capacity_factor":162},142.7,0.7,{"gpu_model":164,"gpu_model_id":165,"gpu_count":31,"total_vram_gb":166,"hour_price":167,"monthly_cost":168,"providers":169,"bandwidth_gbps":170,"fp16_tflops":171,"interconnect_score":154,"parallel_efficiency":172,"score":173,"tier":174,"perf_metric":157,"perf_label":175,"perf_value":176,"workload":10,"cost_metric":177},"RTX 5070",28,48,0.0101,29.49,[148],672,61,0.6,0.5752,"cheapest","98.2 tok\u002Fs",98.2,{"effective_value":178,"effective_unit":157,"capacity_factor":162},68.7,{"gpu_model":180,"gpu_model_id":181,"gpu_count":41,"total_vram_gb":182,"hour_price":183,"monthly_cost":184,"providers":185,"bandwidth_gbps":186,"fp16_tflops":187,"interconnect_score":154,"parallel_efficiency":188,"score":189,"tier":156,"perf_metric":157,"perf_label":190,"perf_value":191,"workload":10,"cost_metric":192},"H200 SXM",11,1128,3.5,20440,[151],4800,989,0.4,0.2197,"934.9 tok\u002Fs",934.9,{"effective_value":193,"effective_unit":157,"capacity_factor":162},654.4,{"gpu_model":195,"gpu_model_id":196,"gpu_count":37,"total_vram_gb":166,"hour_price":197,"monthly_cost":198,"providers":199,"bandwidth_gbps":200,"fp16_tflops":201,"interconnect_score":154,"parallel_efficiency":37,"score":202,"tier":174,"perf_metric":157,"perf_label":203,"perf_value":204,"workload":10,"cost_metric":205},"RTX A6000",6,0.2874,209.8,[148,151],768,77,0.6554,"46.7 tok\u002Fs",46.7,{"effective_value":206,"effective_unit":157,"capacity_factor":162},32.7,[],false,22,"value",[212,216,224,231,238],{"name":195,"vram_gb":166,"cheapest_hour":197,"cost":213,"provider":148,"estimated_tps":204},{"hourly":197,"daily":214,"monthly":198,"yearly":215},6.9,2517.62,{"name":217,"vram_gb":166,"cheapest_hour":218,"cost":219,"provider":148,"estimated_tps":223},"L40S",0.4004,{"hourly":218,"daily":220,"monthly":221,"yearly":222},9.61,292.29,3507.5,52.6,{"name":225,"vram_gb":166,"cheapest_hour":226,"cost":227,"provider":148,"estimated_tps":38},"RTX PRO 6000 WS",0.6689,{"hourly":226,"daily":228,"monthly":229,"yearly":230},16.05,488.3,5859.56,{"name":232,"vram_gb":166,"cheapest_hour":233,"cost":234,"provider":148,"estimated_tps":38},"RTX PRO 6000 S",0.7339,{"hourly":233,"daily":235,"monthly":236,"yearly":237},17.61,535.75,6428.96,{"name":239,"vram_gb":144,"cheapest_hour":240,"cost":241,"provider":148,"estimated_tps":245},"A100 80GB",1.2,{"hourly":240,"daily":242,"monthly":243,"yearly":244},28.8,876,10512,124.1,{"posts_count":38,"benchmarks_count":38}]