DeepSeek R1¶

What it is¶

DeepSeek R1 is a SOTA open-weights reasoning model that uses reinforcement learning (RL) to achieve high-level performance in math, coding, and logical reasoning, rivaling proprietary models like OpenAI's o1.

What problem it solves¶

It provides a high-reasoning, "thinking" model that is accessible as open weights, allowing for complex multi-step planning and validation without the high cost or privacy concerns of proprietary reasoning APIs.

Where it fits in the stack¶

Category: Tool / AI Assistants & Knowledge

Typical use cases¶

Complex code generation and debugging.
Solving advanced mathematical and logical puzzles.
Reasoning-heavy agentic workflows (e.g., architecture planning).

Strengths¶

SOTA performance for an open-weights model.
Transparent "Chain of Thought" (CoT) generation.
Highly cost-effective compared to proprietary reasoning models.

Limitations¶

Higher latency due to the "thinking" process.
May over-reason on simple tasks.

When to use it¶

For tasks where correctness is more important than speed.
When you need a logic-heavy model that you can self-host or access via low-cost providers.

Licensing and cost¶

License: DeepSeek License (Open Weights)
Cost: Low-cost API access via DeepSeek or OpenRouter; free to self-host.

Getting started¶

Installation¶

DeepSeek R1 is accessible via an OpenAI-compatible API or can be run locally using Ollama.

# To run locally via Ollama
ollama run deepseek-r1

Basic API setup (Python)¶

pip install openai

CLI examples¶

1. Basic Chat Completion (curl)¶

curl https://api.deepseek.com/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $DEEPSEEK_API_KEY" \
  -d '{
        "model": "deepseek-v4-pro",
        "messages": [
          {"role": "user", "content": "Explain quantum entanglement like I am five."}
        ],
        "stream": false
      }'

2. Reasoning Mode (Thinking enabled)¶

curl https://api.deepseek.com/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $DEEPSEEK_API_KEY" \
  -d '{
        "model": "deepseek-v4-pro",
        "messages": [
          {"role": "user", "content": "Find the derivative of x^2 * sin(x)."}
        ],
        "thinking": {"type": "enabled"},
        "reasoning_effort": "high"
      }'

3. List Available Models¶

curl https://api.deepseek.com/models \
  -H "Authorization: Bearer $DEEPSEEK_API_KEY"

API examples¶

Python (OpenAI SDK)¶

from openai import OpenAI

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

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[
        {"role": "user", "content": "Write a Python function to sort a list of dictionaries by a specific key."}
    ],
    extra_body={"thinking": {"type": "enabled"}}
)

print(f"Thought: {response.choices[0].message.reasoning_content}")
print(f"Answer: {response.choices[0].message.content}")

Sources / References¶

Contribution Metadata¶

Last reviewed: 2026-05-28
Confidence: high