OpenAI

The GPT models are the best in class LLMs and used as the default LLM by Agents. OpenAI supports a variety of world-class models. See their models here. We recommend experimenting to find the best-suited model for your use-case. Here are some general recommendations:

gpt-4o is good for most general use-cases.
gpt-4o-mini model is good for smaller tasks and faster inference.
o1 models are good for complex reasoning and multi-step tasks.
o3-mini is a strong reasoning model with support for tool-calling and structured outputs, but at a much lower cost.

OpenAI have tier based rate limits. See the docs for more information.

Authentication

Set your OPENAI_API_KEY environment variable. You can get one from OpenAI here.

export OPENAI_API_KEY=sk-***

Example

Use OpenAIChat with your Agent:

from agno.agent import Agent, RunResponse
from agno.models.openai import OpenAIChat

agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    markdown=True
)

# Print the response in the terminal
agent.print_response("Share a 2 sentence horror story.")

View more examples here.

Params

For more information, please refer to the OpenAI docs as well.

Name	Type	Default	Description
`id`	`str`	`"gpt-4o"`	The id of the OpenAI model to use.
`name`	`str`	`"OpenAIChat"`	The name of this chat model instance.
`provider`	`str`	`"OpenAI " + id`	The provider of the model.
`store`	`Optional[bool]`	`None`	Whether or not to store the output of this chat completion request for use in the model distillation or evals products.
`metadata`	`Optional[Dict[str, Any]]`	`None`	Additional metadata to include with the request.
`frequency_penalty`	`Optional[float]`	`None`	Penalizes new tokens based on their frequency in the text so far.
`logit_bias`	`Optional[Any]`	`None`	Modifies the likelihood of specified tokens appearing in the completion.
`logprobs`	`Optional[bool]`	`None`	Include the log probabilities on the logprobs most likely tokens.
`max_tokens`	`Optional[int]`	`None`	The maximum number of tokens to generate in the chat completion.
`max_completion_tokens`	`Optional[int]`	`None`	The maximum number of tokens to generate in completions.
`modalities`	`Optional[List[str]]`	`None`	List of modalities supported by the model.
`audio`	`Optional[Dict[str, Any]]`	`None`	Audio-specific parameters for the model.
`presence_penalty`	`Optional[float]`	`None`	Penalizes new tokens based on whether they appear in the text so far.
`response_format`	`Optional[Any]`	`None`	An object specifying the format that the model must output.
`seed`	`Optional[int]`	`None`	A seed for deterministic sampling.
`stop`	`Optional[Union[str, List[str]]]`	`None`	Up to 4 sequences where the API will stop generating further tokens.
`temperature`	`Optional[float]`	`None`	Controls randomness in the model's output.
`top_logprobs`	`Optional[int]`	`None`	How many log probability results to return per token.
`user`	`Optional[str]`	`None`	A unique identifier representing your end-user.
`top_p`	`Optional[float]`	`None`	Controls diversity via nucleus sampling.
`extra_headers`	`Optional[Any]`	`None`	Additional headers to send with the request.
`extra_query`	`Optional[Any]`	`None`	Additional query parameters to send with the request.
`request_params`	`Optional[Dict[str, Any]]`	`None`	Additional parameters to include in the request.
`api_key`	`Optional[str]`	`None`	The API key for authenticating with OpenAI.
`organization`	`Optional[str]`	`None`	The organization to use for API requests.
`base_url`	`Optional[Union[str, httpx.URL]]`	`None`	The base URL for API requests.
`timeout`	`Optional[float]`	`None`	The timeout for API requests.
`max_retries`	`Optional[int]`	`None`	The maximum number of retries for failed requests.
`default_headers`	`Optional[Any]`	`None`	Default headers to include in all requests.
`default_query`	`Optional[Any]`	`None`	Default query parameters to include in all requests.
`http_client`	`Optional[httpx.Client]`	`None`	An optional pre-configured HTTP client.
`client_params`	`Optional[Dict[str, Any]]`	`None`	Additional parameters for client configuration.
`client`	`Optional[OpenAIClient]`	`None`	The OpenAI client instance.
`async_client`	`Optional[AsyncOpenAIClient]`	`None`	The asynchronous OpenAI client instance.
`structured_outputs`	`bool`	`False`	Whether to use the structured outputs from the Model.
`supports_structured_outputs`	`bool`	`True`	Whether the Model supports structured outputs.
`add_images_to_message_content`	`bool`	`True`	Whether to add images to the message content.
`override_system_role`	`bool`	`True`	Whether to override the system role.
`system_message_role`	`str`	`"developer"`	The role to map the system message to.

OpenAIChat is a subclass of the Model class and has access to the same params.

Introduction

Concepts

Other

How to

Authentication

Example

Params

Introduction

Concepts

Other

How to

Documentation Index

​Authentication

​Example

​Params

Authentication

Example

Params