Metraipolis › archive › dc_d9ddfa1654dc
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression — arXiv:2609.19969
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.19969
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T03:48:09.607Z
- SHA-256 of the archived text
0a552776ac77d1148fc3b5fcd826758f20300a82d9c02423cacf1b9e2a056c76- Stable URL of this page
https://metraipolis.com/documento/dc_d9ddfa1654dc
How to cite
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression — arXiv:2609.19969 — arXiv (en), https://arxiv.org/abs/2609.19969 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d9ddfa1654dc (Metraipolis archive dc_d9ddfa1654dc, sha256 0a552776ac77d1148fc3b5fcd826758f20300a82d9c02423cacf1b9e2a056c76)
BibTeX
@misc{dcd9ddfa1654dc,
title = {DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression — arXiv:2609.19969},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.19969},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d9ddfa1654dc; sha256 0a552776ac77d1148fc3b5fcd826758f20300a82d9c02423cacf1b9e2a056c76},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19969. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation, prefill remains computationally expensive, and large KV caches continue to strain HBM and SSD capacity and data-transfer bandwidth. Together, these compute, storage, and bandwidth demands constitute the primary bottleneck to further lowering deployment costs. To address this challenge, we introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. With its Causal Encoder-Decoder (CED) architecture, the model activates 16B parameters per token during decode but only 8B parameters during prefill, substantially improving cost efficiency for agentic workloads. To push the limits of KV cache compression, DeepSeek-V4.1-Flash combines cross-layer KV cache reuse in Compressed Sparse Attention 2 (CSA2) with FP4 KV caching. These designs reduce its global KV cache footprint (always in HBM) to 890 bytes per token, roughly 1/4 of the corresponding footprint of DeepSeek-V4-Flash. Further, through a dedicated deployment optimization known as SWA Bounded Replay, DeepSeek-V4.1-Flash reduces its persistent KV cache footprint (always on SSD or in host memory) to roughly 1/8 of that of DeepSeek-V4-Flash. Despite its much smaller KV cache footprint, the model delivers substantially better performance than the baseline. In addition, we streamline the DeepSeek-V4 architecture and introduce several efficient architectural extensions. We pretrain DeepSeek-V4.1-Flash on a multimodal corpus comprising 45T tokens and conduct comprehensive post-training, yielding strong performance across diverse text-based and multimodal agentic scenarios. Model checkpoints are available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.19969 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-10T03:24:08.147Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).