[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-replit-code-3b":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":135,"fitting_gpus":203,"community":239},351,"replit-code-3b","Replit Code 3B","replit\u002Freplit-code-v1_5-3b","replit","transformer",null,3300000000,"3.3B",4096,"MIT",130,316,"published","https:\u002F\u002Fhuggingface.co\u002Freplit\u002Freplit-code-v1_5-3b","2026-08-12T03:11:24+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",0.0035,[],[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":79,"INT8":92,"AWQ":96,"GPTQ":109,"GGUF":122},{"minimum":57,"production":65,"recommended":72},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":59,"kv_cache_gb":42,"vram_gb":60,"ram_gb":61,"disk_gb":62,"requirement_source":63,"confidence":64},"minimum",6.15,12.73,32,48.59,"estimated",0.95,{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":59,"kv_cache_gb":68,"vram_gb":69,"ram_gb":70,"disk_gb":71,"requirement_source":63,"confidence":64},"production",16384,64,126.08,189.12,113.59,{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":59,"kv_cache_gb":41,"vram_gb":76,"ram_gb":77,"disk_gb":78,"requirement_source":63,"confidence":64},"recommended",8192,2,28.83,43.24,74.59,{"minimum":80,"production":84,"recommended":88},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":81,"kv_cache_gb":42,"vram_gb":82,"ram_gb":61,"disk_gb":83,"requirement_source":63,"confidence":64},3.07,8.84,43.79,{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":81,"kv_cache_gb":68,"vram_gb":85,"ram_gb":86,"disk_gb":87,"requirement_source":63,"confidence":64},121.84,182.76,108.79,{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":81,"kv_cache_gb":41,"vram_gb":89,"ram_gb":90,"disk_gb":91,"requirement_source":63,"confidence":64},24.76,37.15,69.79,{"minimum":93,"production":94,"recommended":95},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":81,"kv_cache_gb":42,"vram_gb":82,"ram_gb":61,"disk_gb":83,"requirement_source":63,"confidence":64},{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":81,"kv_cache_gb":68,"vram_gb":85,"ram_gb":86,"disk_gb":87,"requirement_source":63,"confidence":64},{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":81,"kv_cache_gb":41,"vram_gb":89,"ram_gb":90,"disk_gb":91,"requirement_source":63,"confidence":64},{"minimum":97,"production":101,"recommended":105},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":98,"kv_cache_gb":42,"vram_gb":99,"ram_gb":61,"disk_gb":100,"requirement_source":63,"confidence":64},1.59,6.97,41.49,{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":98,"kv_cache_gb":68,"vram_gb":102,"ram_gb":103,"disk_gb":104,"requirement_source":63,"confidence":64},119.8,179.7,106.49,{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":98,"kv_cache_gb":41,"vram_gb":106,"ram_gb":107,"disk_gb":108,"requirement_source":63,"confidence":64},22.81,34.21,67.49,{"minimum":110,"production":114,"recommended":118},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":111,"kv_cache_gb":42,"vram_gb":112,"ram_gb":61,"disk_gb":113,"requirement_source":63,"confidence":64},1.61,6.99,41.52,{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":111,"kv_cache_gb":68,"vram_gb":115,"ram_gb":116,"disk_gb":117,"requirement_source":63,"confidence":64},119.83,179.74,106.52,{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":111,"kv_cache_gb":41,"vram_gb":119,"ram_gb":120,"disk_gb":121,"requirement_source":63,"confidence":64},22.83,34.25,67.52,{"minimum":123,"production":127,"recommended":131},{"scenario":58,"context_length":13,"batch_size":37,"weight_gb":124,"kv_cache_gb":42,"vram_gb":125,"ram_gb":61,"disk_gb":126,"requirement_source":63,"confidence":64},1.65,7.04,41.58,{"scenario":66,"context_length":67,"batch_size":31,"weight_gb":124,"kv_cache_gb":68,"vram_gb":128,"ram_gb":129,"disk_gb":130,"requirement_source":63,"confidence":64},119.88,179.82,106.58,{"scenario":73,"context_length":74,"batch_size":75,"weight_gb":124,"kv_cache_gb":41,"vram_gb":132,"ram_gb":133,"disk_gb":134,"requirement_source":63,"confidence":64},22.88,34.33,67.58,{"recommended":136,"cheapest":154,"performance":169,"min_complexity":186,"cluster_options":200,"no_match":201,"required_vram_gb":106,"candidate_count":157,"usage":202},{"gpu_model":137,"gpu_model_id":75,"gpu_count":37,"total_vram_gb":61,"hour_price":138,"monthly_cost":139,"providers":140,"bandwidth_gbps":143,"fp16_tflops":144,"interconnect_score":145,"parallel_efficiency":37,"score":146,"tier":147,"perf_metric":148,"perf_label":149,"perf_value":150,"workload":10,"cost_metric":151},"RTX 5090",0.337,246.01,[141,142],"vast","runpod",1792,209,40,0.6878,"cheapest","tok\u002Fs","1127 tok\u002Fs",1127,{"effective_value":152,"effective_unit":148,"capacity_factor":153},788.9,0.7,{"gpu_model":155,"gpu_model_id":156,"gpu_count":75,"total_vram_gb":157,"hour_price":158,"monthly_cost":159,"providers":160,"bandwidth_gbps":161,"fp16_tflops":162,"interconnect_score":145,"parallel_efficiency":163,"score":164,"tier":147,"perf_metric":148,"perf_label":165,"perf_value":166,"workload":10,"cost_metric":167},"RTX 5070",28,24,0.0101,14.75,[141],672,61,0.8,0.6015,"676.2 tok\u002Fs",676.2,{"effective_value":168,"effective_unit":148,"capacity_factor":153},473.3,{"gpu_model":170,"gpu_model_id":171,"gpu_count":41,"total_vram_gb":172,"hour_price":173,"monthly_cost":174,"providers":175,"bandwidth_gbps":177,"fp16_tflops":178,"interconnect_score":145,"parallel_efficiency":179,"score":180,"tier":181,"perf_metric":148,"perf_label":182,"perf_value":183,"workload":10,"cost_metric":184},"H200 SXM",11,1128,3.5,20440,[176],"lambda",4800,989,0.4,0.2197,"balanced","9660.4 tok\u002Fs",9660.4,{"effective_value":185,"effective_unit":148,"capacity_factor":153},6762.3,{"gpu_model":187,"gpu_model_id":188,"gpu_count":37,"total_vram_gb":157,"hour_price":189,"monthly_cost":190,"providers":191,"bandwidth_gbps":193,"fp16_tflops":194,"interconnect_score":145,"parallel_efficiency":37,"score":195,"tier":147,"perf_metric":148,"perf_label":196,"perf_value":197,"workload":10,"cost_metric":198},"RTX 3090",3,0.0678,49.49,[141,192],"tensordock",936,71,0.6423,"588.7 tok\u002Fs",588.7,{"effective_value":199,"effective_unit":148,"capacity_factor":153},412.1,[],false,"value",[204,208,216,223,231],{"name":187,"vram_gb":157,"cheapest_hour":189,"cost":205,"provider":141,"estimated_tps":197},{"hourly":189,"daily":206,"monthly":190,"yearly":207},1.63,593.93,{"name":209,"vram_gb":157,"cheapest_hour":210,"cost":211,"provider":141,"estimated_tps":215},"RTX 4090",0.1344,{"hourly":210,"daily":212,"monthly":213,"yearly":214},3.23,98.11,1177.34,634,{"name":217,"vram_gb":157,"cheapest_hour":218,"cost":219,"provider":141,"estimated_tps":215},"RTX 4090D",0.1602,{"hourly":218,"daily":220,"monthly":221,"yearly":222},3.84,116.95,1403.35,{"name":224,"vram_gb":157,"cheapest_hour":225,"cost":226,"provider":141,"estimated_tps":230},"A10",0.2015,{"hourly":225,"daily":227,"monthly":228,"yearly":229},4.84,147.1,1765.14,377.4,{"name":232,"vram_gb":157,"cheapest_hour":233,"cost":234,"provider":141,"estimated_tps":238},"L4",0.2022,{"hourly":233,"daily":235,"monthly":236,"yearly":237},4.85,147.61,1771.27,188.7,{"posts_count":38,"benchmarks_count":38}]