DeepSeek

DeepSeek R1 671B

Name: DeepSeek R1 671B
Author: DeepSeek

Frontier

HuggingFace

Ollama

5.2MDownloads13.4KCurtidasJan 2025Publicado131K tokensContextoMITLicença84 ForteQualidade

DeepSeek R1 671B (671B parameters) requires approximately 470.4 GB of VRAM with Q4_K_M quantization. As a Mixture of Experts model with 37B active parameters, it uses less memory than its total parameter count suggests. For the best balance of quality and speed, we recommend hardware with at least 541 GB of VRAM.

Comece agora

— copie e cole para rodar localmente

Copy-paste commands to run DeepSeek R1 671B on your machine.

Run

ollama run deepseek-r1

Quick specs

Parameters671B (37B active)

Architecturemoe (MoE)

Context131K tokens

Modalitytext

Min RAM261.7 GB

Rec. RAM409.3 GB (Q4_K_M)

LicenseMIT

FamilyDeepSeek

✓ Chat✓ Reasoning

About this model

We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as a preliminary step, demonstrated remarkable performance on reasoning. With RL, DeepSeek-R1-Zero naturally emerged with numerous powerful and interesting reasoning behaviors. However, DeepSeek-R1-Zero encounters challenges such as endless repetition, poor readability, and language mixing.

•We directly apply reinforcement learning (RL) to the base model without relying on supervised fine-tuning (SFT) as a preliminary step. This...
•We introduce our pipeline to develop DeepSeek-R1. The pipeline incorporates two RL stages aimed at discovering improved reasoning patterns and...

Modelos relacionados

Opções de quantização

Estimativas de VRAM por nível de quantização

No hardware detected — fit column shows raw VRAM estimates

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	261.7 GB	Low	—
Q3_K_S	3	328.8 GB	Low	—
NVFP4	4	375.8 GB	Medium	—
Q4_K_M	4	409.3 GB	Medium	—
Q5_K_M	5	483.1 GB	High	—
Q6_K	6	550.2 GB	High	—
Q8_0	8	718.0 GB	Very High	—
F16	16	1375.6 GB	Maximum	—

Quality benchmarks

DeepSeek R1 671B benchmark scores

Benchmark verified

Coding

SWE-bench Verified49.2%

HumanEval+85.0%

Aider Polyglot53.3%

LiveCodeBench65.9%

Reasoning

MMLU-Pro84.0%

GPQA Diamond71.5%

MATH-50097.3%

ARC Challenge—

General

Chatbot Arena—

IFEval83.3%

Source: official · 2025-01-20

Compatibilidade de hardware

Estimativas de compatibilidade para todo o hardware

Abrir calculadora

Computing compatibility...

Detalhamento de memória

Reference: RTX 2060 6GB

Weights409.3 GB

KV Cache59.6 GB

Runtime0.9 GB

Headroom0.6 GB

Perguntas frequentes

FAQ — DeepSeek R1 671B

Veja também

Guia de Quantização Metodologia de Pontuação Abrir calculadora