AI
Deepseek just did the impossibleAI Search18 Sept 2026
A separate 168B-parameter memory module that sits in server RAM rather than GPU memory and stores static facts. This frees GPU memory for reasoning.
DeepSeek research on a separate large conditional memory module for static facts, held in server RAM instead of GPU memory.
Updated EU status checked Last shown
Engram at a glance
- What is Engram?
- Engram is a research from DeepSeek. DeepSeek research on a separate large conditional memory module for static facts, held in server RAM instead of GPU memory.
- Who makes Engram?
- Engram is made by DeepSeek.
- Does Engram work in the EU?
- Unverified. No explicit evidence found. Last verified 14 Nov 2026.
- How much does Engram cost?
- Pricing unknown.
- Where was Engram shown?
- In 1 creator video by AI Search. Most recently by AI Search on 18 Sept 2026, at 23:46 in “Deepseek just did the impossible”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
About
Engram is a DeepSeek research project describing a memory module that is separate from the main model's reasoning layers. According to a YouTube creator summary, it is a roughly 168B-parameter component that stores static facts and sits in server RAM rather than GPU memory, which would free GPU memory for reasoning.
The provided official page text is only the DeepSeek landing page, which does not describe Engram. Details such as architecture, benchmarks, release date and licensing could not be confirmed from the supplied material, and the figures above come from a single secondary source.
No standalone product, API or downloadable release for Engram was identified in the provided sources. It should be treated as research rather than a generally available tool.
Key features
- Separate memory module for static facts
- Reported ~168B parameters (per creator summary)
- Runs from server RAM instead of GPU memory
- Aims to free GPU memory for reasoning
1 moment
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Keep exploring
Similar AI tools
Closest neighbours by meaning, not just by category.
DE
DeepSeek
Model · DeepSeek
Chinese AI lab behind the open-weight DeepSeek-V4 and DeepSeek-R1 models, with a chat app and low-cost API.
UnverifiedLanguage Models
Release date n/aFreemium
DV
DeepSeek V4 Flash Vision Experimental
Model · DeepSeek
Experimental vision-enabled version of DeepSeek V4 Flash with image understanding; open weights under MIT, now retired on the API.
Available in EULanguage Models
31 Aug 2026Open source
DE
DeepInfer
Library · DeepSeek
Open-source inference framework from DeepSeek for serving DeepSeek models on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs.
Available in EUPlatforms & Infra
16 Oct 2026Open source
DV
DeepSeek V4 Flash
Model · DeepSeek
Open-weights Flash-tier model in DeepSeek's V4 family, offered via chat and a low-cost API.
Available in EULanguage Models
Release date n/aAPI
DV
DeepSeek V4.1 Flash
Model · DeepSeek
Open-weights multimodal DeepSeek model in the V4 family, offered via API, chat and downloadable weights.
Available in EULanguage Models
Release date n/aOpen source
DS
DSpark
Research · DeepSeek
DeepSeek research component for multi-token output, intended to speed up LLM text generation.
UnverifiedResearch
Release date n/a