Skip to content

DeepSeek R1 70B

by DeepSeek · deepseek-r1 family

70B

parameters

text-generation code-generation reasoning math creative-writing summarization

DeepSeek R1 70B is the largest distilled reasoning model in the DeepSeek R1 series, based on the Llama 3.3 70B architecture. It captures the most reasoning capability from the full DeepSeek R1 671B model through distillation, delivering exceptional performance on complex reasoning tasks. This model approaches the reasoning quality of the full R1 model on many benchmarks while requiring far less compute. It excels at advanced mathematics, competitive programming, scientific reasoning, and complex analytical tasks. Multi-GPU setups are recommended for comfortable inference.

Quick Start with Ollama

ollama run 70b-q4_K_M
Resources Ollama Hugging Face Official Page Research Paper
Creator DeepSeek
Parameters 70B
Architecture transformer-decoder
Context 128K tokens
Released Jan 20, 2025
License MIT
Ollama deepseek-r1:70b

Quantization Options

Format File Size VRAM Required Quality Ollama Tag
Q4_K_M rec 34.9 GB 43.5 GB 70b-q4_K_M
Q5_K_M 40.8 GB 50.5 GB 70b-q5_K_M
Q8_0 64.8 GB 72 GB 70b-q8_0

Compatible Hardware

Q4_K_M requires 43.5 GB VRAM

Compatible Hardware

HardwareVRAMTypeFitEst. Speed
Mac Studio M4 Ultra 512GB512 GBmacRuns~19 tok/s
Mac Pro M2 Ultra 192GB192 GBmacRuns~18 tok/s
Mac Studio M4 Ultra 192GB192 GBmacRuns~19 tok/s
Mac Studio M4 Max 128GB128 GBmacRuns~13 tok/s
MacBook Pro M4 Max 128GB128 GBmacRuns~13 tok/s
MacBook Pro M5 Max 128GB128 GBmacRuns~13 tok/s
NVIDIA RTX PRO 6000 Blackwell96 GBgpuRuns~44 tok/s
MacBook Pro M3 Max 96GB96 GBmacRuns~9 tok/s
Mac mini M4 Pro 64GB64 GBmacRuns~6 tok/s
Mac Studio M4 Max 64GB64 GBmacRuns~13 tok/s
MacBook Pro M4 Max 64GB64 GBmacRuns~13 tok/s
MacBook Pro M5 Max 64GB64 GBmacRuns~13 tok/s
NVIDIA RTX 6000 Ada Generation48 GBgpuRuns (tight)~22 tok/s
NVIDIA RTX A600048 GBgpuRuns (tight)~18 tok/s
NVIDIA RTX PRO 5000 Blackwell48 GBgpuRuns (tight)~22 tok/s
Mac mini M4 Pro 48GB48 GBmacRuns (tight)~6 tok/s
MacBook Pro M3 Max 48GB48 GBmacRuns (tight)~9 tok/s
MacBook Pro M4 Pro 48GB48 GBmacRuns (tight)~6 tok/s
MacBook Pro M4 Max 48GB48 GBmacRuns (tight)~13 tok/s
MacBook Pro M5 Max 48GB48 GBmacRuns (tight)~9 tok/s
MacBook Pro M5 Pro 48GB48 GBmacRuns (tight)~6 tok/s
Mac Studio M4 Max 36GB36 GBmacCPU Offload~4 tok/s
MacBook Pro M3 Pro 36GB36 GBmacCPU Offload~1 tok/s
MacBook Pro M5 Max 36GB36 GBmacCPU Offload~3 tok/s
NVIDIA RTX 5000 Ada Generation32 GBgpuCPU Offload~5 tok/s
NVIDIA GeForce RTX 509032 GBgpuCPU Offload~12 tok/s
iMac M4 32GB32 GBmacCPU Offload~1 tok/s
Mac mini M4 32GB32 GBmacCPU Offload~1 tok/s
MacBook Air M4 32GB32 GBmacCPU Offload~1 tok/s
MacBook Air M5 32GB32 GBmacCPU Offload~1 tok/s
MacBook Pro M5 32GB32 GBmacCPU Offload~1 tok/s
76 hardware device(s) cannot run this model at Q4_K_M.

Benchmark Scores

85.5
mmlu