Getting the model to think
DeepSeek V4-Pro reasons by default! For the chat-style tasks you've done so far, the model returns a clean final answer; for deeper tasks like data analysis, code debugging, math, or research, it works through the problem step-by-step before answering.
In this exercise, you'll send V4-Pro a typical math problem that might be presented to high-school math students and let its default reasoning do the work.
> math_problem
Calculate the area of a house in m^2 comprised of two rooms of sizes 3mx2m and 4mx5m.
This exercise is part of the course
Working with DeepSeek in Python
Exercise instructions
- Create a request to send the
math_problemprompt to the"deepseek-ai/DeepSeek-V4-Pro"model via the Chat Completions endpoint.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
client = OpenAI(api_key="", base_url="https://api.together.xyz/v1")
# Create a request to the model — reasoning runs by default
response = client.chat.completions.____(
model="deepseek-ai/____",
max_tokens=400,
messages=[
{"role": "user",
"content": ____}]
)
print(response.choices[0].message.content)