deepseek-ai
/

DeepSeek-V4-Pro

Text Generation

8-bit precision

Model card Files Files and versions

Instructions to use deepseek-ai/DeepSeek-V4-Pro with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use deepseek-ai/DeepSeek-V4-Pro with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="deepseek-ai/DeepSeek-V4-Pro")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-V4-Pro")
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-V4-Pro")

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps

How to use deepseek-ai/DeepSeek-V4-Pro with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "deepseek-ai/DeepSeek-V4-Pro"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4-Pro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/deepseek-ai/DeepSeek-V4-Pro

How to use deepseek-ai/DeepSeek-V4-Pro with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "deepseek-ai/DeepSeek-V4-Pro" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4-Pro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "deepseek-ai/DeepSeek-V4-Pro" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "deepseek-ai/DeepSeek-V4-Pro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use deepseek-ai/DeepSeek-V4-Pro with Docker Model Runner:
```
docker model run hf.co/deepseek-ai/DeepSeek-V4-Pro
```

Resources

View closed (8)

Upload wms-pro-dashboard-template (1).zip

#178 opened about 16 hours ago by

Upload 11 files

#177 opened about 16 hours ago by

越用越觉得好用

#176 opened 2 days ago by

Fixed a type annotation.

#175 opened 4 days ago by

finance

#174 opened 5 days ago by

Add Claw-Eval evaluation results

#173 opened 5 days ago by

Question about causal safety in DeepSeek-V4 CSA prefill retrieval

#172 opened 6 days ago by

Update README.md

#170 opened 7 days ago by

Too much positivity bias

#169 opened 10 days ago by

Thank You, DeepSeek! Empower Local Open Source with Smaller Models for Consumer Hardware/感谢 DeepSeek！赋能本地开源，社区强烈呼吁推出适合家用硬件的小型模型 (

#168 opened 12 days ago by

Deepseek web issue

#167 opened 13 days ago by

Performance feedback

#166 opened 14 days ago by

Add YC-Bench benchmark result (avg $1,066,426)

#165 opened 15 days ago by

Will there be small models like 12b?

#164 opened 15 days ago by

Assigned weights for different teachers. 教师权重的分配

#162 opened 16 days ago by

Where is HCA implemented?

#161 opened 16 days ago by

思考链下会出现英文混乱

#160 opened 16 days ago by

Partial Rotary Positional Embedding 的笔误？

#159 opened 17 days ago by

模型文件注释里的形状似乎写错了

#158 opened 18 days ago by

什么时候支持api上传文档就好了

#156 opened 19 days ago by

deepseek-ai/DeepSeek-V4-Pro

#153 opened 20 days ago by

16 - 24B models with FP8 quantization

#152 opened 20 days ago by

Support other low-bit quantization methods?

#151 opened 21 days ago by

源神启动！恭喜deepseekV4顺利上线，合影！

#150 opened 21 days ago by

Общение

#149 opened 21 days ago by

Update README.md to add SGLang deployment option

#148 opened 22 days ago by

Add pipeline tag to allow for HF Inference Providers

#147 opened 22 days ago by

Add chat template

#146 opened 22 days ago by

Instruction following still not good

#145 opened 22 days ago by

核弹椅子摔瘫坐

#144 opened 22 days ago by

Is it possible to have API output with origin completion prompt including | DSML | token

#143 opened 22 days ago by

恭喜deepseek v4可以撑地了

#142 opened 22 days ago by

代码能力真的是世界第一吗！！！

#141 opened 22 days ago by

开源的模型可以自己去评测，闭源的模型怎么测试呢？最近扎堆发布，GPT5.5，HY3.0 Preview，大家是不是内部都知道各个产品的时间呢？

#140 opened 22 days ago by

恭喜！

#139 opened 22 days ago by

恭喜上线

#138 opened 22 days ago by

GOOD JOB!

#136 opened 22 days ago by

恭喜恭喜

#135 opened 22 days ago by

deepseek niu bee plus

#134 opened 22 days ago by

V4必火！

#133 opened 22 days ago by

火钳流明

#132 opened 22 days ago by

I'm very disappointed in this release

#131 opened 22 days ago by

DeepSeek 666

#130 opened 22 days ago by

Technical Report Summary

#129 opened 22 days ago by

Legendary drop

#128 opened 22 days ago by

合影留念~~~

#127 opened 22 days ago by

关于 "Observations and Proposals" 中激活函数建议的疑问：去掉 gate projection 为何能放宽 EP 带宽要求？

#126 opened 22 days ago by

Help me recover my deepseek account

#125 opened 22 days ago by

请问是否有计划出自己的coding plan？

#124 opened 22 days ago by

V4留念 ~~~

#123 opened 22 days ago by