Implementation:Langchain ai Langchain OpenAI Model Profiles

Knowledge Sources	Langchain_ai_Langchain
Domains	LLM, OpenAI, Model Configuration
Last Updated	2026-02-11 00:00 GMT

Overview

OpenAI Model Profiles is an auto-generated data module that defines capability profiles (token limits, modality support, tool calling, structured output) for OpenAI models.

Description

The _profiles.py file in the langchain-openai partner package contains a dictionary named _PROFILES that maps OpenAI model identifiers to their capability metadata. Each profile entry includes maximum input/output token counts, supported input/output modalities (text, image, audio, video), reasoning output capability, tool calling and structured output support, and additional flags for image URL inputs, PDF inputs, PDF tool messages, image tool messages, and tool choice support. This data is auto-generated by the langchain-profiles CLI tool from the models.dev project and should not be manually edited. The profiles are consumed by BaseChatOpenAI and AzureChatOpenAI to automatically set model profile metadata.

Usage

This module is used internally by OpenAI chat model classes during initialization to populate model profiles with appropriate token limits and capability flags. It is not typically imported directly by end users.

Code Reference

Source Location

• Repository: Langchain_ai_Langchain
• File: libs/partners/openai/langchain_openai/data/_profiles.py
• Lines: 1-832

Signature

_PROFILES: dict[str, dict[str, Any]] = {
    "model-name": {
        "max_input_tokens": int,
        "max_output_tokens": int,
        "text_inputs": bool,
        "image_inputs": bool,
        "audio_inputs": bool,
        "video_inputs": bool,
        "text_outputs": bool,
        "image_outputs": bool,
        "audio_outputs": bool,
        "video_outputs": bool,
        "reasoning_output": bool,
        "tool_calling": bool,
        "structured_output": bool,
        "image_url_inputs": bool,
        "pdf_inputs": bool,
        "pdf_tool_message": bool,
        "image_tool_message": bool,
        "tool_choice": bool,
    },
    ...
}

Import

from langchain_openai.data._profiles import _PROFILES

I/O Contract

Inputs

Name	Type	Required	Description
(none)	--	--	This is a static data module with no runtime inputs.

Outputs

Name	Type	Description
_PROFILES	dict[str, dict[str, Any]]	A dictionary mapping model name strings to profile dictionaries containing token limits, modality flags, and capability flags.

Registered Models

The module includes profiles for the following OpenAI models (among others):

Model	Max Input Tokens	Max Output Tokens	Image Input	Reasoning	Tool Calling	Structured Output
gpt-4o	128,000	16,384	Yes	No	Yes	Yes
gpt-4o-mini	128,000	16,384	Yes	No	Yes	Yes
gpt-4.1	1,047,576	32,768	Yes	No	Yes	Yes
gpt-4.1-mini	1,047,576	32,768	Yes	No	Yes	Yes
gpt-4.1-nano	1,047,576	32,768	Yes	No	Yes	Yes
gpt-5	272,000	128,000	Yes	Yes	Yes	Yes
gpt-5-mini	272,000	128,000	Yes	Yes	Yes	Yes
o1	200,000	100,000	Yes	Yes	Yes	Yes
o3	200,000	100,000	Yes	Yes	Yes	Yes
o3-mini	200,000	100,000	No	Yes	Yes	Yes
o4-mini	200,000	100,000	Yes	Yes	Yes	Yes
gpt-4	8,192	8,192	No	No	Yes	No
gpt-3.5-turbo	16,385	4,096	No	No	No	No
text-embedding-3-small	8,191	1,536	No	No	No	--
text-embedding-3-large	8,191	3,072	No	No	No	--

Usage Examples

Accessing a Model Profile

from langchain_openai.data._profiles import _PROFILES

profile = _PROFILES.get("gpt-4o")
print(profile["max_input_tokens"])    # 128000
print(profile["tool_calling"])        # True
print(profile["structured_output"])   # True
print(profile["image_inputs"])        # True

Related Pages

• Environment:Langchain_ai_Langchain_OpenAI_API_Credentials