OpenAI Provider - Python AI SDK

Overview

The OpenAI provider offers comprehensive support for all OpenAI services including GPT models, embeddings, function calling, and structured output. It’s the most feature-complete provider in the SDK.

Quick Start

from ai_sdk import openai, generate_text

model = openai("gpt-4.1-mini")
res = generate_text(model=model, prompt="Hello, world!")
print(res.text)

Available Models

Chat Models

Model	Description	Max Tokens	Input / Cached Input / Output (per 1M tokens)
`gpt-4.1`	GPT-4.1 (latest)	128k	$2.00 /$ 0.50 / $8.00
`gpt-4.1-mini`	GPT-4.1 Mini	128k	$0.40 /$ 0.10 / $1.60
`gpt-4.1-nano`	GPT-4.1 Nano	128k	$0.10 /$ 0.025 / $0.40
`gpt-4.5-preview`	GPT-4.5 Preview	128k	$75.00 /$ 37.50 / $150.00
`gpt-4o`	GPT-4o (Omni)	128k	$2.50 /$ 1.25 / $10.00
`gpt-4o-mini`	GPT-4o Mini	128k	$0.15 /$ 0.075 / $0.60
`gpt-4o-mini-audio-preview`	GPT-4o Mini Audio Preview	128k	$0.15 / - /$ 0.60
`o3`	OpenAI O3	128k	$2.00 /$ 0.50 / $8.00
`o3-deep-research`	O3 Deep Research	128k	$10.00 /$ 2.50 / $40.00
`o3-pro`	O3 Pro	128k	$20.00 / - /$ 80.00
`o3-mini`	O3 Mini	128k	$1.10 /$ 0.55 / $4.40
`o4-mini`	O4 Mini	128k	$1.10 /$ 0.275 / $4.40
`o4-mini-deep-research`	O4 Mini Deep Research	128k	$2.00 /$ 0.50 / $8.00
`o1`	OpenAI O1	128k	$15.00 /$ 7.50 / $60.00
`o1-pro`	O1 Pro	128k	$150.00 / - /$ 600.00
`o1-mini`	O1 Mini	128k	$1.10 /$ 0.55 / $4.40
`codex-mini-latest`	Codex Mini (latest)	128k	$1.50 /$ 0.375 / $6.00
`gpt-4o-mini-search-preview`	GPT-4o Mini Search Preview	128k	$0.15 / - /$ 0.60
`gpt-4o-search-preview`	GPT-4o Search Preview	128k	$2.50 / - /$ 10.00
`computer-use-preview`	Computer Use Preview	128k	$3.00 / - /$ 12.00

| gpt-4o-mini | Latest GPT-4 model | 128k |

5/1M input,

15/1M output |

Embedding Models

Model	Dimensions	Cost
`text-embedding-3-large`	3072	$0.13/1M tokens
`text-embedding-3-small`	1536	$0.02/1M tokens
`text-embedding-ada-002`	1536	$0.0001/1K tokens

Basic Usage

Text Generation

from ai_sdk import openai, generate_text

model = openai("gpt-4.1-mini")
res = generate_text(
    model=model,
    prompt="Write a haiku about programming"
)
print(res.text)

Streaming

import asyncio
from ai_sdk import openai, stream_text

async def main():
    model = openai("gpt-4.1-mini")
    stream_res = stream_text(
        model=model,
        prompt="Tell me a story about a robot"
    )

    async for chunk in stream_res.text_stream:
        print(chunk, end="", flush=True)

asyncio.run(main())

With System Instructions

from ai_sdk import openai, generate_text

model = openai("gpt-4.1-mini")
res = generate_text(
    model=model,
    system="You are a helpful coding assistant. Always provide clear, concise explanations.",
    prompt="Explain what recursion is in simple terms"
)
print(res.text)

Advanced Features

Structured Output

OpenAI supports native structured output with response_format="json_object":

from ai_sdk import openai, generate_object
from pydantic import BaseModel

class User(BaseModel):
    name: str
    age: int
    email: str

model = openai("gpt-4.1-mini")
res = generate_object(
    model=model,
    schema=User,
    prompt="Create a user profile for John Doe, age 30"
)
print(res.object)  # User(name='John Doe', age=30, email='john@example.com')

Function Calling

from ai_sdk import openai, generate_text, tool

def get_weather(city: str) -> str:
    """Get weather for a city."""
    weather_data = {
        "New York": "72°F, Sunny",
        "London": "55°F, Rainy"
    }
    return weather_data.get(city, "Weather data not available")

weather_tool = tool(
    name="get_weather",
    description="Get current weather for a city",
    parameters={
        "type": "object",
        "properties": {
            "city": {"type": "string", "description": "City name"}
        },
        "required": ["city"]
    },
    execute=get_weather
)

model = openai("gpt-4.1-mini")
res = generate_text(
    model=model,
    prompt="What's the weather like in New York?",
    tools=[weather_tool]
)
print(res.text)

Embeddings

from ai_sdk import openai, embed_many, cosine_similarity

# Create embedding model
embed_model = openai.embedding("text-embedding-3-small")

# Embed multiple texts
texts = [
    "The cat sat on the mat.",
    "A dog was lying on the rug.",
    "Python is a programming language."
]

result = embed_many(model=embed_model, values=texts)

# Calculate similarity
similarity = cosine_similarity(result.embeddings[0], result.embeddings[1])
print(f"Similarity: {similarity:.3f}")

Vision Models

from ai_sdk import openai, generate_text
from ai_sdk.types import CoreUserMessage, ImagePart

model = openai("gpt-4.1")  # Vision-capable model

# Create message with image
message = CoreUserMessage(content=[
    ImagePart(image_url="https://example.com/image.jpg"),
    "Describe this image in detail."
])

res = generate_text(model=model, messages=[message])
print(res.text)

Configuration

API Key

Set your OpenAI API key:

export OPENAI_API_KEY="sk-..."

Or pass it directly:

model = openai("gpt-4.1-mini", api_key="sk-...")

Default Parameters

Configure default parameters for all requests:

model = openai(
    "gpt-4.1-mini",
    temperature=0.7,
    max_tokens=1000,
    top_p=0.9,
    frequency_penalty=0.1,
    presence_penalty=0.1,
    user="my-app/123"  # For analytics
)

Organization

For team accounts:

model = openai(
    "gpt-4.1-mini",
    organization="org-..."  # Your organization ID
)

Parameters

Common Parameters

Parameter	Type	Default	Description
`model`	`str`	-	Model identifier (required)
`api_key`	`str`	`None`	API key (uses OPENAI_API_KEY env var)
`organization`	`str`	`None`	Organization ID
`base_url`	`str`	`https://api.openai.com/v1`	API base URL

Generation Parameters

Parameter	Type	Default	Description
`temperature`	`float`	`1.0`	Controls randomness (0.0 = deterministic)
`max_tokens`	`int`	`None`	Maximum tokens to generate
`top_p`	`float`	`1.0`	Nucleus sampling parameter
`frequency_penalty`	`float`	`0.0`	Reduces repetition
`presence_penalty`	`float`	`0.0`	Encourages new topics
`response_format`	`str`	`None`	`"json_object"` for structured output
`seed`	`int`	`None`	For reproducible results

Embedding Parameters

Parameter	Type	Default	Description
`encoding_format`	`str`	`"float"`	`"float"` or `"base64"`
`dimensions`	`int`	`None`	Output dimensions (model-specific)

Error Handling

Rate Limiting

import time
from ai_sdk import openai, generate_text

def generate_with_retry(prompt, max_retries=3):
    model = openai("gpt-4.1-mini")

    for attempt in range(max_retries):
        try:
            return generate_text(model=model, prompt=prompt)
        except Exception as e:
            if "rate_limit" in str(e).lower() and attempt < max_retries - 1:
                wait_time = 2 ** attempt
                print(f"Rate limited, waiting {wait_time}s...")
                time.sleep(wait_time)
                continue
            raise

res = generate_with_retry("Hello!")

Token Limits

from ai_sdk import openai, generate_text

def truncate_prompt(prompt, max_tokens=4000):
    """Truncate prompt to fit within token limits."""
    # Rough estimation: 1 token ≈ 4 characters
    max_chars = max_tokens * 4
    if len(prompt) > max_chars:
        return prompt[:max_chars] + "..."
    return prompt

model = openai("gpt-4.1-mini")
long_prompt = "A very long prompt..." * 1000
truncated = truncate_prompt(long_prompt)
res = generate_text(model=model, prompt=truncated)

Best Practices

1. Model Selection

Choose the right model for your use case:

# For simple tasks - fast and cheap
model = openai("gpt-3.5-turbo")

# For complex reasoning - powerful but expensive
model = openai("gpt-4.1")

# For structured output - native JSON support
model = openai("gpt-4.1-mini")

2. Cost Optimization

Monitor and optimize costs:

from ai_sdk import openai, generate_text

model = openai("gpt-4.1-mini")
res = generate_text(model=model, prompt="Hello!")

if res.usage:
    input_cost = res.usage.prompt_tokens * 0.00000015  # $0.15/1M tokens
    output_cost = res.usage.completion_tokens * 0.0000006  # $0.6/1M tokens
    total_cost = input_cost + output_cost
    print(f"Cost: ${total_cost:.6f}")

3. Prompt Engineering

Use clear, specific prompts:

# Good
prompt = """
You are a helpful coding assistant. The user will ask you questions about Python programming.

Please provide:
1. A clear explanation
2. A code example
3. Best practices to follow

User question: {user_question}
"""

# Avoid
prompt = "Help me with Python"

4. Streaming for Long Responses

Use streaming for better user experience:

import asyncio
from ai_sdk import openai, stream_text

async def generate_long_response():
    model = openai("gpt-4.1-mini")
    stream_res = stream_text(
        model=model,
        prompt="Write a detailed tutorial about Python decorators"
    )

    print("Generating response...")
    async for chunk in stream_res.text_stream:
        print(chunk, end="", flush=True)
    print("\nDone!")

asyncio.run(generate_long_response())

5. Structured Output for Reliability

Use structured output for consistent results:

from ai_sdk import openai, generate_object
from pydantic import BaseModel
from typing import List

class CodeReview(BaseModel):
    issues: List[str]
    suggestions: List[str]
    score: int

model = openai("gpt-4.1-mini")
res = generate_object(
    model=model,
    schema=CodeReview,
    prompt="Review this Python code: def hello(): print('world')"
)

review = res.object
print(f"Score: {review.score}/10")
print(f"Issues: {review.issues}")

Troubleshooting

Common Issues

Invalid API Key
Error: Invalid API key
- Check your API key is correct
- Ensure you have sufficient credits
Model Not Found
Error: The model `gpt-4.1-mini` does not exist
- Verify model name spelling
- Check if model is available in your region
Rate Limiting
Error: Rate limit exceeded
- Implement exponential backoff
- Consider upgrading your plan
Token Limit Exceeded
Error: Request too large
- Reduce input length
- Use a model with higher token limits

Debug Mode

Enable detailed logging:

import logging
logging.basicConfig(level=logging.DEBUG)

from ai_sdk import openai, generate_text
model = openai("gpt-4.1-mini")
res = generate_text(model=model, prompt="Hello!")

OpenAI models are constantly being updated. Check the OpenAI API documentation for the latest model availability and pricing.

​Overview

​Quick Start

​Available Models

​Chat Models

​Embedding Models

​Basic Usage

​Text Generation

​Streaming

​With System Instructions

​Advanced Features

​Structured Output

​Function Calling

​Embeddings

​Vision Models

​Configuration

​API Key

​Default Parameters

​Organization

​Parameters

​Common Parameters

​Generation Parameters

​Embedding Parameters

​Error Handling

​Rate Limiting

​Token Limits

​Best Practices

​1. Model Selection

​2. Cost Optimization

​3. Prompt Engineering

​4. Streaming for Long Responses

​5. Structured Output for Reliability

​Troubleshooting

​Common Issues

​Debug Mode

Overview

Quick Start

Available Models

Chat Models

Embedding Models

Basic Usage

Text Generation

Streaming

With System Instructions

Advanced Features

Structured Output

Function Calling

Embeddings

Vision Models

Configuration

API Key

Default Parameters

Organization

Parameters

Common Parameters

Generation Parameters

Embedding Parameters

Error Handling

Rate Limiting

Token Limits

Best Practices

1. Model Selection

2. Cost Optimization

3. Prompt Engineering

4. Streaming for Long Responses

5. Structured Output for Reliability

Troubleshooting

Common Issues

Debug Mode