Developing story · 2 episodes · 10 Sept to 18 Sept
DeepSeek Launches Open Long Context Agent Model
DeepSeek has introduced the V4.1-Flash model, featuring a one-million token context window and an MIT license to support long-running agents. The release positions the company as a provider of cost-effective, open-source AI solutions for extended autonomous tasks.
-
Engram offloading boosts DeepSeek-V4.1-Flash performance on B300 by up to 1.6x
The SemiAnalysis report describes Engram, a token‑embedding extension that stores recurring patterns in a lookup table. By moving the Engram table—about 24 rows per layer, roughly 12.4 KiB per token…
1 source -
DeepSeek releases V4.1-Flash with 1M context and MIT license
DeepSeek AI has released DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts model designed to address the memory bottlenecks of long-horizon agent workloads. The model features a 552B parameter…
6 sources primary source
Who and what
DeepseekOpenDesign ArenaOpenAINVIDIAAMDGoogle CloudMicrosoft AzureOracleMetaV4.1-FlashDeepSeek-V4.1-FlashGPT-6 Astra