DeepSeek's R1 and V3 models delivered a major shock to the AI industry when they matched frontier closed-source models on reasoning benchmarks — and released weights anyone can download. In 2026 they remain among the strongest open-weight options for math, coding, and structured reasoning. This guide covers everything you need to run them on your own machine, from hardware requirements to context tuning.
What changed in 2026
- Distilled models became the practical choice. DeepSeek released R1 distilled variants (1.5B through 70B) trained to preserve the reasoning ability of the full 671B MoE while running on consumer hardware. Most users should start here.
- Ollama added native DeepSeek support.
ollama pull deepseek-r1:8b is now a single-command install with sensible defaults. No manual GGUF conversion needed.
- DeepSeek V3 (685B MoE) weights shipped. V3 is the general-purpose counterpart to R1's reasoning focus. At full size it requires 4–8 A100 GPUs; the community-quantised Q4 version runs on 2× A100 80 GB.
- Privacy concerns are real. DeepSeek is a Chinese company; their hosted API sends data to Chinese servers. Running locally eliminates this concern entirely — which is why local deployment surged.
Hardware requirements by variant
| Variant |
Size (Q4) |
Min RAM/VRAM |
Recommended |
Speed (approx.) |
| R1 1.5B Q4 |
~1 GB |
4 GB |
Any modern laptop |
100+ t/s |
| R1 7B Q4 |
~5 GB |
8 GB |
16 GB RAM / M2 |
50–70 t/s |
| R1 8B Q4 |
~5 GB |
8 GB |
16 GB RAM / M2 |
50–70 t/s |
| R1 14B Q4 |
~9 GB |
12 GB |
M3 Pro / RTX 3080 |
35–50 t/s |
| R1 32B Q4 |
~20 GB |
24 GB |
M3 Max / RTX 4090 |
20–30 t/s |
| R1 70B Q4 |
~42 GB |
48 GB |
M4 Max / 2× RTX 4090 |
12–18 t/s |
| V3 (full, Q4) |
~350 GB |
Multi-GPU server |
4× A100 80 GB |
Server-dependent |
Unified memory on Apple Silicon counts fully toward model loading.
Setup with Ollama (recommended)
Step 1: Install Ollama from ollama.com — available for macOS, Linux, and Windows (WSL).
Step 2: Pull the model variant that fits your hardware:
ollama pull deepseek-r1:8b
Other tags: deepseek-r1:14b, deepseek-r1:32b, deepseek-r1:70b
Step 3: Run interactively:
ollama run deepseek-r1:8b
Step 4: Set context length (critical — default is too short):
Create a Modelfile:
FROM deepseek-r1:8b
PARAMETER num_ctx 16384
Then: ollama create deepseek-r1-ctx16k -f Modelfile
Step 5: Use the OpenAI-compatible API at http://localhost:11434/v1 — drop-in compatible with any OpenAI SDK call.
Setup with llama.cpp (advanced)
Download the GGUF from Hugging Face (search bartowski/DeepSeek-R1-Distill-Llama-8B-GGUF), then:
./llama-server -m deepseek-r1-8b-q4_k_m.gguf \
--ctx-size 16384 --n-gpu-layers 99 --port 8080
--n-gpu-layers 99 offloads all layers to GPU; reduce if you have less VRAM.
How to pick the right variant
- Have 8–16 GB RAM? Use R1 7B or 8B. Strong for most tasks; runs fast.
- Have a 24 GB GPU or M3 Pro? Use R1 14B — meaningful quality jump over 8B.
- Have 48 GB+ unified memory or a 24 GB GPU pair? R1 32B–70B for serious reasoning work.
- Need general chat vs. deep reasoning? R1 is the reasoning model; V3 is the general model. For most users, R1 distilled beats V3 distilled on coding and math.
Common mistakes
Using the default context window. Ollama defaults to 2048 tokens for many models. DeepSeek R1 benefits enormously from 8k–16k contexts — always set num_ctx explicitly.
Running the full 671B MoE without the right hardware. The full model requires proper MoE-aware serving (not just enough RAM). Without it, performance is abysmal. Use distilled variants on consumer hardware.
Ignoring the thinking tokens. DeepSeek R1 emits <think>...</think> blocks before its answer. These are useful for debugging but count toward your context. Strip them if you are piping output programmatically.
Not checking temperature. R1's chain-of-thought reasoning benefits from lower temperature (0.1–0.6) for consistent outputs. Default 0.8 introduces more variance in reasoning paths.
Downloading the wrong quantisation. Q2_K is too lossy for reasoning tasks. Q4_K_M is the minimum for reliable R1 performance. Q6_K or Q8_0 is better if you have the VRAM headroom.
What to skip
- DeepSeek hosted API if privacy or data residency is a concern — the whole point of running locally is to keep data on-prem.
- R1 1.5B for any task requiring real reasoning. It is useful for classification and short completions but degrades severely on multi-step problems.
- Mixing R1 and V3 distills randomly. They have different strengths: use R1 distills for reasoning-heavy tasks, V3 distills for general instruction-following and chat.
FAQ
Is DeepSeek R1 safe to run locally?
The model weights are open and auditable. Running locally means no data is sent to DeepSeek servers. However, like all open models, fine-tuning can alter behaviour — use official distilled releases for predictable output.
How does R1 compare to Llama 3.3 70B?
On math and coding benchmarks, R1 70B generally scores higher than Llama 3.3 70B. For general instruction-following and conversation, Llama 3.3 is often preferred. R1 shines when chain-of-thought reasoning matters.
Can I use DeepSeek R1 for RAG or agent workflows?
Yes. Ollama exposes an OpenAI-compatible API, so any framework (LangChain, LlamaIndex, LangGraph) works without modification. Tool calling is supported in the R1 distilled variants.
Does running locally require a GPU?
No. Apple Silicon Macs use unified memory and run all layers efficiently via Metal. On Windows/Linux, CPU-only inference is possible but slow (5–15 t/s for 8B); a GPU is strongly recommended for practical use.
Where to go next
Best local AI models in 2026 compares DeepSeek against Llama, Qwen, Phi, and Mistral across hardware tiers. AI coding agents ranked in 2026 shows how to wire local models into coding agent workflows. Best AI agent builders in 2026 covers the frameworks that integrate with local model servers.