The Frontier · Tags

#KV cache

1 published article

• Compute & power · The Record

  DeepSeek's 17 September V4.1-Flash report cuts global KV cache to 890 bytes per token and leaves the quality cost unmeasured

  The 51-page arXiv paper pairs FP4 KV caching, cross-layer cache reuse and an encoder-decoder split. It reports no serving speed and no ablation for its two riskiest shortcuts.