LOCAL_AI_STACK
v2.2 · 2026-06-12
← back to models

DeepSeek-R1-0528

deepseek-r1-0528-full
REASONING

DeepSeek-R1-0528 is a 671B reasoning model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.

PROVIDER
DeepSeek
FAMILY
DeepSeek-R1
MODEL TYPE
reasoning
PARAMETERS
671B (37B active)
MODALITIES
text
ARCHITECTURE
Mixture-of-Experts reasoning model
CONTEXT WINDOW
128,000 tokens
TRAINING TOKENS
RELEASE DATE
2025-05-28
LICENSE & LINKS
SETUP HINTS
OLLAMA
ollama run deepseek-r1:671b
LLAMA.CPP
Use a cited GGUF build with current llama.cpp-compatible tooling.
HARDWARE FIT
CPU only (no GPU)
no
8 GB RAM
unknown
16 GB RAM
unknown
32 GB RAM
unknown
8 GB VRAM
unknown
12 GB VRAM
unknown
24 GB VRAM
unknown
Apple Silicon (unified memory)
unknown
Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.
BEST FOR
·Explaining the original R1 reasoning model and separating cloud/server use from local distills for beginners.
AVOID IF
·You expect an ordinary Mac, gaming PC, or small GPU to run it well locally.
CAVEATS
·Not practical for normal local users.
·Use distilled 8B, 14B, or 32B variants instead.
·Long-context and full-weight deployment require specialized server infrastructure.
·Censorship and political-bias caveats still matter.
·Commercial-use status from the supplied research: Allowed under MIT license.
·Beginner summary from supplied research: Full DeepSeek-R1-0528 is a server-class reasoning reference model, not a beginner local download. List it only to explain the source teacher model behind distills and to set expectations about hardware, speed, and censorship caveats clearly.
SOURCE URLS
FIELD EVIDENCE
hfGgufRepos
q5FileSizeGb
trainingTokens
LANGUAGES
Unknown
CAPABILITIES
reasoning tuned
DeepSeek-R1-Distill-Qwen-32BDeepSeek-V3.2