[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f22y50ydqmi2q1":3,"$f3rzxi0hr4syhl":26,"$f2aqt8dsqk55hx":28,"$f3bp8zr6m6ceiu":32,"$f8ypa021er9qz":53,"$f3n2dq8bla4rfj":55},{"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},85,"DeepSeek-V4.1-Flash-1017","DeepSeek",763.2,"CHAT","OPEN_SOURCE",null,"https:\u002F\u002Fhuggingface.co\u002Fdeepseek-ai\u002FDeepSeek-V4.1-Flash-1017","DeepSeek-V4.1-Flash 的 10 月 17 日检查点更新（MIT）。相对原版基准全面抬升：HLE 无工具 36.8→42.6、Terminal-Bench 31.2→42.3、DeepSWE 74.2→77.0、Agents Last Exam 31.8→35.4；输出更短使实际速度更快、成本更低。1M 上下文，激活参数 37B，与原版共享权重架构与部署方式。","Checkpoint refresh of DeepSeek-V4.1-Flash released Oct 17 (MIT). Broad benchmark gains over the base model: HLE w\u002Fo tools 36.8→42.6, Terminal-Bench 31.2→42.3, DeepSWE 74.2→77.0, Agents Last Exam 31.8→35.4; shorter outputs mean faster effective speed and lower cost. 1M context, 37B active parameters, same architecture and deployment path as the original.",[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":51,"page":30,"pageSize":52},[34,44],{"id":35,"type":36,"title":37,"excerpt":38,"author":39,"replyCount":41,"createdAt":42,"lastRepliedAt":10,"resolvedAt":10,"editedAt":10,"model":43,"hardware":10,"variant":10,"record":10},38,"HELP","DeepInfer（开源推理引擎，5.8k stars）只支持 V4.1-Flash + SM100\u002FAscend 950——消费卡用户怎么看？","GitHub 上 DeepInfer 这两天很火（Apache-2.0，5.8k stars），号称是 DeepSeek-V4.1-Flash 的专用推理引擎，性能比通用栈高一截。但仔细看了下支持矩阵： 模型：只有 DeepSeek-V4.",{"nickname":40,"avatarUrl":10},"matrix_w",3,"2026-10-20T12:21:10.422Z",{"name":5},{"id":45,"type":36,"title":46,"excerpt":47,"author":48,"replyCount":41,"createdAt":49,"lastRepliedAt":10,"resolvedAt":10,"editedAt":10,"model":50,"hardware":10,"variant":10,"record":10},33,"DeepSeek-V4.1-Flash-1017：从 API 切过去要注意什么？","官方说 1017 是免费 MIT 更新，基准全线上移（HLE +5.8、Terminal-Bench +11.1），输出还更短——理论上同样的调用更便宜。 但部署侧有两个点想确认： 「没有 Jinja chat template、改用 de",{"nickname":40,"avatarUrl":10},"2026-10-19T12:48:53.557Z",{"name":5},2,20,{"items":54},[],{"items":56,"total":30,"page":30,"pageSize":41},[57],{"slug":58,"category":59,"publishedAt":60,"titleZh":61,"titleEn":62,"summaryZh":63,"summaryEn":64,"models":65,"hardwares":70},"deepseek-v4-1-flash-1017-2026-10-19","NEWS","2026-10-19T10:00:00.000Z","DeepSeek-V4.1-Flash-1017 发布：免费 MIT 更新，推理更快、输出更短","DeepSeek-V4.1-Flash-1017 ships: free MIT update, faster reasoning, shorter outputs","10 月 17 日 DeepSeek 上线 V4.1-Flash 的 1017 检查点（MIT）。官方基准全面抬升：HLE 36.8→42.6、Terminal-Bench 31.2→42.3、Apex 64.6→72.1；输出更短带来实际更快与更省。本地量化当日铺开，站内模型目录已收录（763B 总参保持目录口径、无单卡实测）。","DeepSeek shipped the 1017 checkpoint of V4.1-Flash on Oct 17 (MIT). Broad official gains: HLE 36.8→42.6, Terminal-Bench 31.2→42.3, Apex 64.6→72.1; shorter outputs mean faster wall-clock and lower cost. Local quants landed same-week; catalog row added on our site (763B total stays catalog-only, no single-GPU runs).",[66,69],{"id":67,"name":68},18,"DeepSeek-V4.1-Flash",{"id":4,"name":5},[]]