[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$ffyfmwg1f21f5":3,"$f1ycxvp12vtxa5":26,"$fzyoa15nfba3x":28,"$f3bp8zd5i9icib":32,"$f8ypa021er9qz":47,"$f29uog2c75lxx6":49},{"id":4,"name":5,"publisher":6,"paramsB":7,"useCase":8,"licenseStatus":9,"licenseUrl":10,"sourceUrl":11,"descriptionZh":12,"descriptionEn":13,"imageUrl":10,"downloadUrls":14,"stats":17,"variants":20,"derivations":21,"measuredHardware":24,"updatedAt":25},86,"Beam-1.0-501B-A23B","Reflection",501,"REASONING","OPEN_SOURCE",null,"https:\u002F\u002Fhuggingface.co\u002Freflection\u002FBeam-1.0-501B-A23B","Reflection 发布的 501B 总参 \u002F 23B 激活 MoE 推理模型（Apache 2.0），主打「推理时更少思考 token」——官方口径约为同级模型的 1\u002F3。1M 上下文，BF16 \u002F FP8 \u002F NVFP4 权重，10 月 16 日起社区 GGUF 量化与 DFlash2 等投机解码后端陆续跟进。","Reflection's 501B-total \u002F 23B-active MoE reasoning model (Apache 2.0) built around a simple claim: about 1\u002F3 the inference-time reasoning tokens of comparable models. 1M context, BF16 \u002F FP8 \u002F NVFP4 weights; community GGUF quants and speculative-decoding backends like DFlash2 have been landing since Oct 16.",[15],{"url":11,"label":16},"HuggingFace",{"recordCount":18,"variantCount":18,"hardwareCount":18,"unlinkedRecordCount":18,"evidence":19},0,{"L0":18,"L1":18,"L2":18},[],{"parents":22,"children":23},[],[],[],"2026-10-19T12:27:43.798Z",{"favorited":27},false,{"items":29,"total":18,"page":30,"pageSize":31},[],1,50,{"items":33,"total":30,"page":30,"pageSize":46},[34],{"id":35,"type":36,"title":37,"excerpt":38,"author":39,"replyCount":41,"createdAt":42,"lastRepliedAt":10,"resolvedAt":10,"editedAt":10,"model":43,"hardware":44,"variant":10,"record":10},34,"HELP","Beam-1.0-501B-A23B 本地到底能不能跑？算了一下显存账","Reflection 新出的 Beam 501B 总参 \u002F 23B 激活，Apache 2.0，官方说推理 token 用量约同级 1\u002F3。社区 GGUF 已经出来了，问题是显存账： IQ1S 按社区量化实测约 122GB 权重 我的 St",{"nickname":40,"avatarUrl":10},"bit_squeezer",3,"2026-10-19T12:48:53.559Z",{"name":5},{"name":45},"AMD Ryzen AI Max+ 395",20,{"items":48},[],{"items":50,"total":30,"page":30,"pageSize":41},[51],{"slug":52,"category":53,"publishedAt":54,"titleZh":55,"titleEn":56,"summaryZh":57,"summaryEn":58,"models":59,"hardwares":70},"nvidia-rtx-spark-launch-2026-10-19","NEWS","2026-10-19T14:00:00.000Z","NVIDIA RTX Spark 上市：$2,399 起、128GB 统一内存，站内首批 59 条实测","NVIDIA RTX Spark launches: from $2,399, 128GB unified memory, first 59 on-site runs","N1X 超级芯片整机（Blackwell RTX GPU + 20 核 CPU），LPDDR5X 24-128GB，笔记本 $2,399 起。StorageReview 实测 STREAM 120GB\u002Fs、Llama-3.1-8B Q4_K_M 56.5 tok\u002Fs。站内硬件收录 + 59 条 L2 记录：llama.cpp Q4_K_M 锚点 8B 54.7 \u002F 35B-A3B 88.7，与实测对齐。Windows 下 128GB 整机 GPU 侧可用约 62.5GB。","N1X superchip system (Blackwell RTX GPU + 20-core CPU), LPDDR5X 24-128GB, laptops from $2,399. StorageReview measured STREAM at 120GB\u002Fs and Llama-3.1-8B Q4_K_M at 56.5 tok\u002Fs. On-site hardware row + 59 L2 records: llama.cpp Q4_K_M anchors land at 54.7 (8B) \u002F 88.7 (35B-A3B), matching the measurement. Under Windows roughly 62.5GB of the 128GB unit is GPU-usable today.",[60,63,66,67],{"id":61,"name":62},47,"Qwen3.6-35B-A3B",{"id":64,"name":65},68,"Llama-3.1-8B-Instruct",{"id":4,"name":5},{"id":68,"name":69},87,"Hy-Memory-30B-A3B",[71],{"id":72,"name":73},110,"NVIDIA RTX Spark"]