DeepSeek Guide 2025: R1 Model, Chat & API Setup

1. What is DeepSeek?

DeepSeek is a Chinese AI research lab (founded 2023, backed by High-Flyer hedge fund) that develops and open-sources large language models. Their key breakthrough: achieving GPT-4o and o1-level performance at a fraction of the training cost, then releasing the weights for free under the MIT license.

2. DeepSeek R1 vs V3: Which to Use?

DeepSeek has two flagship models with very different purposes. Choosing the right one makes a big difference in quality and cost.

3. Using DeepSeek Chat (Free)

The easiest way to try DeepSeek — no account, no API key, no credit card. Just open the browser and start chatting.

4. DeepSeek R1 on Ollama (Local, Free)

Running DeepSeek R1 locally with Ollama gives you complete privacy — no data leaves your machine. It's free after the initial download (the weights are MIT licensed). You'll need at least 8GB RAM for the 7B model, 32GB+ for 70B.

5. DeepSeek API Setup

The DeepSeek API is fully OpenAI-compatible — same request/response format, same SDK. You only need to change two things: the base URL and the model name.

6. Python Code Examples

The DeepSeek API uses the OpenAI Python SDK — just set a different base_url. Install with pip install openai.

import openai

client = openai.OpenAI(
    api_key="YOUR_DEEPSEEK_API_KEY",
    base_url="https://api.deepseek.com"
)

response = client.chat.completions.create(
    model="deepseek-chat",  # DeepSeek V3
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Explain quantum entanglement simply."}
    ]
)
print(response.choices[0].message.content)

# Use DeepSeek R1 for reasoning tasks
response = client.chat.completions.create(
    model="deepseek-reasoner",  # DeepSeek R1
    messages=[
        {"role": "user", "content": "Solve: If a train travels at 80 mph for 2.5 hours, then 60 mph for 1.5 hours, what is the average speed for the whole trip?"}
    ]
)
# R1 returns a reasoning_content field with the chain-of-thought
print(response.choices[0].message.reasoning_content)
print(response.choices[0].message.content)

7. DeepSeek R1 vs OpenAI o1: Comparison

DeepSeek R1 directly competes with OpenAI o1 as a reasoning model. On most benchmarks, they perform similarly — but R1 costs 1/30th the price.

8. Privacy Considerations

FAQ