[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fkad3uimwh8ng":3,"$fi8zcivjofjqa":27,"$f30xyn6gr8dzpx":29,"$f31udvjuf2303k":33,"$f12f9c00yrhx38":36,"$f2xmvtqfs4yaaj":38},{"id":4,"name":5,"publisher":6,"paramsB":7,"useCase":8,"licenseStatus":9,"licenseUrl":10,"sourceUrl":11,"descriptionZh":12,"descriptionEn":13,"imageUrl":14,"downloadUrls":15,"stats":18,"variants":21,"derivations":22,"measuredHardware":25,"updatedAt":26},18,"DeepSeek-V4.1-Flash","DeepSeek",763.2,"CHAT","OPEN_SOURCE","https:\u002F\u002Fhuggingface.co\u002Fdeepseek-ai\u002FDeepSeek-V4.1-Flash\u002Fblob\u002Fmain\u002FLICENSE","https:\u002F\u002Fhuggingface.co\u002Fdeepseek-ai\u002FDeepSeek-V4.1-Flash","DeepSeek 2026-09-10 发布的开权重图文理解模型，763.2B 总参数，MIT 许可。DeepSeek-V4.1 的 Flash 档，重点改进 KV cache 压缩，总参数约 763B（MoE）。","Open-weight vision-language model from DeepSeek, released 2026-09-10, 763.2B total, MIT. The Flash tier of DeepSeek-V4.1, focused on KV-cache compression, roughly 763B total parameters (MoE).","\u002Fuploads\u002Fmodels\u002Fm18-dcb49907717f.png",[16],{"url":11,"label":17},"HuggingFace",{"recordCount":19,"variantCount":19,"hardwareCount":19,"unlinkedRecordCount":19,"evidence":20},0,{"L0":19,"L1":19,"L2":19},[],{"parents":23,"children":24},[],[],[],"2026-09-26T04:16:34.202Z",{"favorited":28},false,{"items":30,"total":19,"page":31,"pageSize":32},[],1,50,{"items":34,"total":19,"page":31,"pageSize":35},[],20,{"items":37},[],{"items":39,"total":61,"page":31,"pageSize":62},[40,51],{"slug":41,"category":42,"publishedAt":43,"titleZh":44,"titleEn":45,"summaryZh":46,"summaryEn":47,"models":48,"hardwares":50},"deepseek-v4-1-flash-derivatives","NEWS","2026-10-02T05:10:00.000Z","发布三周，DeepSeek V4.1 Flash 衍生生态成形：GGUF 量化、abliterated 与 FP8 社区版","Three Weeks On, the DeepSeek V4.1 Flash Derivative Ecosystem Takes Shape: GGUF Quants, Abliterated and FP8 Community Builds","得益于 MIT 许可，DeepSeek V4.1 Flash 发布数小时内社区衍生版即上线：abliterated 全精度版、GGUF 本地量化版、FP8 uncensored 版（下载量最高）相继出现，另有 MixedQ2 混合量化与 forgequant 量化工具。官方未参与或背书任何衍生版本。","Thanks to the MIT license, community derivatives of DeepSeek V4.1 Flash appeared within hours of release: a full-precision abliterated build, GGUF quants for local inference, and an FP8 uncensored variant (the most downloaded), plus a MixedQ2 mixed-quant build and the forgequant quantization toolkit. DeepSeek has not created or endorsed any derivative.",[49],{"id":4,"name":5},[],{"slug":52,"category":42,"publishedAt":53,"titleZh":54,"titleEn":55,"summaryZh":56,"summaryEn":57,"models":58,"hardwares":60},"deepseek-v4-1-flash-open","2026-10-01T05:00:00.000Z","DeepSeek V4.1 Flash 开源：每 token KV Cache 仅 890 字节，百万上下文不再昂贵","DeepSeek V4.1 Flash Open-Sourced: 890 Bytes of KV Cache per Token Makes 1M Context Cheap","DeepSeek 发布并开源 V4.1 Flash（MIT 许可）：552B 非对称 MoE，每 token 全局 KV Cache 仅 890 字节、约为上代 1\u002F4，1M 上下文 KV 仅约 890MB；Agent 基准超越 V4 Pro，API 输入 $0.30\u002F百万 token、闲时半价。","DeepSeek has released and open-sourced V4.1 Flash under MIT: a 552B asymmetric MoE whose global KV cache is just 890 bytes per token — a quarter of its predecessor — bringing a 1M-token context down to ~890MB of KV. It beats V4 Pro on agent benchmarks; API input is $0.30 per million tokens with half-price off-peak.",[59],{"id":4,"name":5},[],2,3]