ModelVerificadoExterno

DeepSeek V4 Flash GGUF

TipoOtherPublicador@batiaiCnota

Fonte do modelo

Tipo: Other
Tamanho do arquivo: 10,4–44,7 GB
Atualizado: 25 de set. de 2026
Publicador: @batiai
Parâmetros: 284B
Tarefas: Geração de texto

Trecho da fonte

Artefatos de modelo

24 artefatos

deepseek-ai-DeepSeek-V4-Flash-Q3_K_M-00001-of-00003.gguf

gguf · 43,9 GB · SHA-256 e424fbdafcfe…d1ac

deepseek-ai-DeepSeek-V4-Flash-Q3_K_M-00002-of-00003.gguf

gguf · 44,5 GB · SHA-256 4aab56411e31…e4cf

deepseek-ai-DeepSeek-V4-Flash-Q3_K_M-00003-of-00003.gguf

gguf · 37,7 GB · SHA-256 de1c9d5f9a11…2e8f

deepseek-ai-DeepSeek-V4-Flash-Q4_K_M-00001-of-00004.gguf

gguf · 44,6 GB · SHA-256 63b9ee8944d9…ca44

deepseek-ai-DeepSeek-V4-Flash-Q4_K_M-00002-of-00004.gguf

gguf · 44,7 GB · SHA-256 1b15b5728ad4…a951

deepseek-ai-DeepSeek-V4-Flash-Q4_K_M-00003-of-00004.gguf

gguf · 44,7 GB · SHA-256 3bee8e8ecf25…700d

deepseek-ai-DeepSeek-V4-Flash-Q4_K_M-00004-of-00004.gguf

gguf · 26,1 GB · SHA-256 313eddbdea65…8e81

deepseek-ai-DeepSeek-V4-Flash-Q5_K_M-00001-of-00005.gguf

gguf · 44,6 GB · SHA-256 dc76b13a8daf…1325

deepseek-ai-DeepSeek-V4-Flash-Q5_K_M-00002-of-00005.gguf

gguf · 44,7 GB · SHA-256 c86c95fa725b…3d50

deepseek-ai-DeepSeek-V4-Flash-Q5_K_M-00003-of-00005.gguf

gguf · 44,4 GB · SHA-256 8a9fbcb7e04b…3232

deepseek-ai-DeepSeek-V4-Flash-Q5_K_M-00004-of-00005.gguf

gguf · 43,8 GB · SHA-256 28bbb03ebe67…5177

deepseek-ai-DeepSeek-V4-Flash-Q5_K_M-00005-of-00005.gguf

gguf · 10,4 GB · SHA-256 b95730920e6c…a9f8

deepseek-ai-DeepSeek-V4-Flash-Q6_K-00001-of-00005.gguf

gguf · 44,5 GB · SHA-256 c367da964213…4c71

deepseek-ai-DeepSeek-V4-Flash-Q6_K-00002-of-00005.gguf

gguf · 43,7 GB · SHA-256 f2dbcbdeab2a…abe7

deepseek-ai-DeepSeek-V4-Flash-Q6_K-00003-of-00005.gguf

gguf · 43,7 GB · SHA-256 ddd2cc069cea…dcea

deepseek-ai-DeepSeek-V4-Flash-Q6_K-00004-of-00005.gguf

gguf · 43,6 GB · SHA-256 83423424ce00…213d

deepseek-ai-DeepSeek-V4-Flash-Q6_K-00005-of-00005.gguf

gguf · 41,9 GB · SHA-256 195dc4690805…5bff

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00001-of-00007.gguf

gguf · 44,6 GB · SHA-256 1ad034c5149a…33a6

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00002-of-00007.gguf

gguf · 44,6 GB · SHA-256 130886694cbd…2bd8

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00003-of-00007.gguf

gguf · 44,6 GB · SHA-256 d094782758b1…67ac

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00004-of-00007.gguf

gguf · 44,6 GB · SHA-256 c73f44236246…c3c2

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00005-of-00007.gguf

gguf · 44,6 GB · SHA-256 dda28d1b6137…50d6

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00006-of-00007.gguf

gguf · 44,6 GB · SHA-256 f7ffb8976c8c…ff1b

deepseek-ai-DeepSeek-V4-Flash-Q8_0-00007-of-00007.gguf

gguf · 13,7 GB · SHA-256 1356e06ea390…da28

Trechos de fonte

3 trechos

Readme11 de jul.huggingface.co/batiai/DeepSeek-V4-Flash-GGUF

--- language: - en - zh - ko license: mit tags: - gguf - deepseek - quantized - batiai - 284b - moe - frontier - early-access base_model: deepseek-ai/DeepSeek-V4-Flash pipeline_tag: text-generation library_name: bati.cpp --- # DeepSeek V4-Flash GGUF — BatiAI Early Access <p align="center"> <a href="https://flow.bati.ai"><img src="https://img.shields.io/badge/BatiFlow-macOS%20AI%20Automation-blue?style=for-the-badge&logo=apple"></a> <a href="https://github.com/batiai/bati.cpp"><img src="https://img.shields.io/badge/Inference-bati.cpp-purple?style=for-the-badge"></a> <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"><img src="https://img.shields.io/badge/source-deepseek--ai-orange?style=for-the-badge"></a> </p> > ⚠️ **Early Access — requires `bati.cpp` to run.** DeepSeek V4-Flash was released on > 2026-04-24 and is **not yet supported by `ggml-org/llama.cpp` master**. This GGUF was > converted with [`batiai/bati.cpp`](https://github.com/batiai/bati.cpp) — BatiAI's own > inference library — and inference also requires the same library. **Ollama is not yet > compatible** (will auto-update once mainline merges V4 support). ## Why ship now? V4-Flash is a frontier-class model (284B-A13B, top-tier on SWE-Bench Pro). Mainline support will likely take 1-2 more weeks. We provide this **interim release** so power users can evaluate the model immediately. When mainline merges support, this repo will gain **imatrix-calibrated IQ3_XXS / IQ4_XS / Q5_K_M** plus Ollama push. ## Available Quants | Quant | Size | Recommended hardware | Notes | |-------|------|----------------------|-------| | **Q3_K_M** | **127 GB** | 128 GB unified memory (M4 Max) | 3-bit, smallest | | **Q4_K_M** | **161 GB** | 192 GB+ Mac | balanced — recommended | | **Q5_K_M** | **188 GB** | 256 GB+ Mac | higher fidelity | | **Q6_K** | **218 GB** | 384 GB+ Mac | near-lossless | | **Q8_0** | **282 GB** | M3 Ultra 512 GB | original FP4 → Q8_0 dequant | All quants signed by BatiAI (`general.author=BatiAI`, `general.url=https://flow.bati.ai`). > **Note**: IQ-quants (IQ3_XXS / IQ4_XS) are tracked in `bati.cpp v0.2.0`. They require > imatrix calibration, and `llama-imatrix` currently segfaults during V4-Flash model > context init in the fork. Will be added once that path is fixed (or once mainline > llama.cpp merges V4 support). K-quants above use `bati.cpp v0.1.2`'s integer-...

Note17 de jul.huggingface.co/batiai/DeepSeek-V4-Flash-GGUF

Source context: 6664 downloads · 11 likes · Pipeline text-generation · Library bati.cpp · Repo batiai/DeepSeek-V4-Flash-GGUF

Note11 de jul.huggingface.co/batiai/DeepSeek-V4-Flash-GGUF

Source context: 6727 downloads · 11 likes · Pipeline text-generation · Library bati.cpp · Repo batiai/DeepSeek-V4-Flash-GGUF