Last Updated: November 21, 2025
OpenAI API
GPT integration guide
Core Endpoints
| Item | Description |
|---|---|
Chat Completions
|
GPT-4, GPT-3.5-turbo chat |
Completions
|
Legacy text completion |
Embeddings
|
Text to vectors |
Images
|
DALL-E image generation |
Audio
|
Whisper transcription |
Moderation
|
Content filtering |
Chat Completion
import openai
openai.api_key = "your-api-key"
response = openai.ChatCompletion.create(
model="gpt-4",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Explain quantum computing"}
],
temperature=0.7,
max_tokens=200
)
print(response.choices[0].message.content)
# Streaming response
for chunk in openai.ChatCompletion.create(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "Tell me a joke"}],
stream=True
):
print(chunk.choices[0].delta.get("content", ""), end="")Parameters
| Item | Description |
|---|---|
temperature
|
0-2, higher = more random |
max_tokens
|
Response length limit |
top_p
|
Nucleus sampling (0-1) |
frequency_penalty
|
Reduce repetition |
presence_penalty
|
Encourage new topics |
stop
|
Stop sequences |
Best Practices
- Use system messages to set behavior
- Lower temperature for factual tasks
- Implement token counting for cost control
- Handle rate limits with exponential backoff
💡 Pro Tips
Quick Reference
GPT-4 is more capable but more expensive than GPT-3.5