DeepSeek Cheat Sheet
A quick reference for DeepSeek: the difference between the chat and reasoning models, when to use Deep Think and Search, and how to prompt each one well.
Models & modes
| Feature | Description | Example / tip |
|---|---|---|
| The fast general model for everyday questions, writing and code. | Draft an email, explain a concept | |
| A reasoning model that thinks through hard problems step by step before answering. | Multi-step maths, logic, planning | |
| Turns on the reasoning model for a tougher prompt. | Enable it for a complex problem | |
| Pulls in current information from the web for up-to-date answers. | Turn on for recent facts |
Go further
Ready to master it for real?
This reference is a great start. Go from the basics to pro workflows in our in-depth DeepSeek course.
Where to use it
| Feature | Description | Example / tip |
|---|---|---|
| Chat for free in the browser or the mobile app. | chat.deepseek.com | |
| Call the models from your own apps, with reasoning available too. | Route hard tasks to the reasoner | |
| The models are openly available and can be self-hosted or fine-tuned. | Run a distilled model locally |
Prompting tips
| Feature | Description | Example / tip |
|---|---|---|
| Use the fast chat model for simple work and reasoning mode only when a problem has real steps. | Don't use Deep Think to write a tweet | |
| With R1, state the goal and constraints clearly, then let it work - you don't need to add 'think step by step'. | Give it the full problem and rules up front | |
| As with any assistant, include the background and the exact output shape you want. | Answer as a checklist, max 10 items |
Frequently asked questions
What's the difference between DeepSeek-V3 and DeepSeek-R1?
V3 is the fast, general-purpose chat model for everyday tasks like writing, questions and code. R1 is a reasoning model that works through problems step by step before answering, which makes it stronger on maths, logic and multi-step planning but slower. Switch to it with the Deep Think toggle.
When should I turn on Deep Think?
Use it for problems that genuinely have several steps - maths, debugging tricky logic, planning or analysis. For simple questions and quick writing, the standard chat model is faster and just as good.
Is DeepSeek free?
Yes, DeepSeek is free to use on the web and in the app, and the model weights are openly available. The API is paid but low cost, with reasoning available as a separate model.
Do ChatGPT prompt techniques work in DeepSeek?
Yes. Clear roles, context and format instructions transfer directly. The one difference is that with the R1 reasoning model you can skip 'think step by step' - it already reasons on its own.
More cheat sheets
Just getting started? Try a free course
Beginner-friendly courses in Blender, Photoshop, Figma, Canva and Claude - free.
Go from reference to real skills
Cheat sheets are great for quick lookups. Our in-depth courses take you from the fundamentals to professional-level mastery.
Master DeepSeek with the full course