Get your own DeepSeek-R1
DeepSeek is a popular reasoning model that excels at complex tasks such as math, code, and natural language. This solution shows how to call the DeepSeek-R1 API and how to deploy models of various sizes. Get started in as little as 5 minutes, with no coding required.
Intended customers
Users who require low response latency
Users who need to customize parameters for specific requirements
Users who need high concurrency and large-scale computing power
Solution overview
Diverse solutions to choose from
Call DeepSeek-R1 based on MaaS
Entry-level5-minute deployment
10Recommended
Alibaba Cloud Model Studio
Intended customers
Users who want to quickly try out or integrate the full version model
Users who need to quickly build AI agents with low code
Benefits
Ready for use immediately after activation
Supports the full version of DeepSeek-R1
Deploy the distilled version of DeepSeek-R1 based on PaaS
Easy10-minute deployment
9
Platform for AI
Intended customers
Users who want to deploy independently but have limited Operations and Maintenance (O&M) capabilities
Users who need support for high concurrency and high-performance computing
Benefits
All models in the DeepSeek series
One-click deployment with multiple acceleration frameworks, such as BladeLLM, SGLang, and vLLM
Cost-effective, disaster-tolerant, and highly available distributed inference services
Efficient cost control: Spot instances can reduce costs by up to 90%
Full version API call
Call the full version of DeepSeek-R1 based on MaaS
Call the API using Alibaba Cloud Model Studio
2868889
Call DeepSeek-R1 or other models provided by Model Studio. Model Studio provides a standardized API, eliminating the need to build your own infrastructure. It also includes load balancing and automatic scaling to ensure stable API calls. Chatbox can further simplify the process, allowing you to configure and use the models through a graphical interface instead of the command line.
5 minutes
$0.1
Alibaba Cloud Model Studio
Distilled version model deployment
Deploy the distilled version of DeepSeek-R1 based on PaaS
Deploy using Platform for AI
2868642
This solution deploys DeepSeek-R1-Distill-Qwen-7B with one click from the Model Gallery of PAI. PAI supports accelerated deployment with frameworks such as BladeLLM, SGLang, and vLLM. For the full version DeepSeek-R1, PAI also offers a cost-effective, highly available, and disaster-tolerant distributed deployment method. If you want to maintain your own models, PAI-DSW and PAI-DLC support fine-tuning of DeepSeek models. This solution is ideal for users who require one-click deployment, inference acceleration, and support for high concurrency.
10 minutes
$1.18 (This cost is estimated based on the instance type used in the example and running the resources for one hour or less. Actual costs may vary depending on the instance types you use. The final price is subject to the quote displayed on the console and the final bill.)
Platform for AI (PAI)
More recommendations