DeepSeek/Mixture of Experts

DeepSeek-R1 684.5B

Name: DeepSeek-R1 684.5B
Author: DeepSeek

reasoningchatThinkingDistilled

684.5B

Parameters (37B active)

160K

Context length

Benchmarks

Quantizations

1.6M

HF downloads

Architecture

MoE

Released

2025-01-20

Layers

KV Heads

128

Head Dim

Family

deepseek

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_XXS	2.38	204.1 GB	low
IQ2_M	2.93	251.2 GB	low
Q2_K	3.16	270.9 GB	low
IQ3_XXS	3.25	278.6 GB	low
IQ3_XS	3.5	300.0 GB	low
Q3_K_S	3.64	311.9 GB	low
IQ3_M	3.76	322.2 GB	low
Q3_K_M	4	342.7 GB	low
Q3_K_L	4.3	368.4 GB	moderate
IQ4_XS	4.46	382.1 GB	moderate
Q4_K_S	4.67	400.1 GB	moderate
Q4_K_M	4.89	418.9 GB	good
Q5_K_S	5.57	477.1 GB	good
Q5_K_M	5.7	488.2 GB	good
Q6_K	6.56	561.8 GB	excellent
Q8_0	8.5	727.8 GB	lossless
FP16	16	1369.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN DEEPSEEK-R1 684.5B NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (2)

Arena Elo1373

BigCodeBench50.0

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run deepseek-r1:684.5b-instruct-q4_K_M

Downloads and runs automatically. Add --verbose for speed stats.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace Ollama Library GGUF Downloads Build Hardware

Find the best GPU for DeepSeek-R1 684.5B

Build Hardware for DeepSeek-R1 684.5B

Model Card

View on HuggingFace

Read the full model card for detailed information about this model.