3 Ways to Use the Full DeepSeek R1 Reasoning Model & Third-Party Services
Due to the explosive popularity of DeepSeek AI during the Spring Festival, a surge in user traffic and possible large-scale malicious attacks have caused DeepSeek's servers to be frequently busy, making it impossible to use DeepSeek normally. Below is a summary of three ways to access the full-strength DeepSeek R1 reasoning model:
1. Local Deployment:
Install DeepSeek via Ollama or LM Studio — LM Studio is recommended as it is simpler. You can also deploy it using these two tools to a machine you don't use often, and then set up external network access.
2. Third-Party Services:
A collection of third-party services offering the full-strength DeepSeek R1 reasoning model:
Metaso AI Search: https://metaso.cn
SiliconFlow: https://cloud.siliconflow.cn/i/g8UM7ldN (Recommended)
Baidu Cloud Qianfan: https://console.bce.baidu.com/qianfan/modelcenter/model/model/buildin/list
ByteDance Volcano Engine: https://console.volcengine.com/ark/region:ark+cn-beijing/axperience
Alibaba Cloud Bailian: https://api.together.ai/playground/chat/deepseek-ai/deepseek-r1
SCNet: https://chat.scnet.cn/#/home
360 Nano AI Search: https://n.cn
Fireworks: https://fireworks.ai/models/fireworks/deepseek-r1
The above services are freely accessible within China:
Github: https://github.com/marketplace/mopart/azureml-depseek/deepseek-r1/playground
POE: https://poe.com/depsek-r1
Cursor: https://cursor.sh
Monica: https://monica.im
Lambda: https://lambdalabs.com
Cerebras: https://cerebras.ai
Perplexity: https://Perplexity.ai
3. Cloud Deployment:
Within China, you can choose Alibaba Cloud or Tencent Cloud; overseas, choose Azure.
The above are the three ways to use the full-strength DeepSeek R1 reasoning model. Choose based on your actual situation.
Script Planner Assistant▸
AI Apps
AI News 