Implementation:Langchain ai Langchain XAI Model Profiles

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

Overview

Auto-generated model profile data for xAI's Grok family of models, defining capabilities and token limits used by ChatXAI.

Description

_PROFILES is a module-level dictionary in the langchain-xai partner package that contains auto-generated model profile metadata for all supported xAI Grok models. Each profile maps a model name to a dictionary of capability flags and token limits (max input/output tokens, supported input/output modalities, reasoning output, and tool calling support). This data is derived from the models.dev project and is consumed by ChatXAI to populate ModelProfile objects for LangChain's model profile registry.

This file is auto-generated by the langchain-profiles CLI tool and should not be edited manually.

Usage

This module is imported internally by ChatXAI during model initialization. It is not intended for direct external use, but can be referenced to understand the capabilities of supported xAI models.

Code Reference

Source Location

• Repository: Langchain_ai_Langchain
• File: libs/partners/xai/langchain_xai/data/_profiles.py
• Lines: 1-327

Signature

_PROFILES: dict[str, dict[str, Any]] = {
    "grok-4-fast-non-reasoning": { ... },
    "grok-3-fast": { ... },
    "grok-4": { ... },
    ...
}

Import

from langchain_xai.data._profiles import _PROFILES

I/O Contract

Inputs

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

Outputs

Name	Type	Description
_PROFILES	dict[str, dict[str, Any]]	Dictionary mapping model names to their capability profiles.

Supported Models

Model Name	Max Input Tokens	Max Output Tokens	Image Inputs	Reasoning Output	Tool Calling
grok-4	256,000	64,000	No	Yes	Yes
grok-4-fast	2,000,000	30,000	Yes	Yes	Yes
grok-4-fast-non-reasoning	2,000,000	30,000	Yes	No	Yes
grok-4-1-fast	2,000,000	30,000	Yes	Yes	Yes
grok-4-1-fast-non-reasoning	2,000,000	30,000	Yes	No	Yes
grok-3	131,072	8,192	No	No	Yes
grok-3-fast	131,072	8,192	No	No	Yes
grok-3-mini	131,072	8,192	No	Yes	Yes
grok-3-mini-fast	131,072	8,192	No	Yes	Yes
grok-2	131,072	8,192	No	No	Yes
grok-2-vision	8,192	4,096	Yes	No	Yes
grok-code-fast-1	256,000	10,000	No	Yes	Yes
grok-beta	131,072	4,096	No	No	Yes
grok-vision-beta	8,192	4,096	Yes	No	Yes

Profile Schema

Each profile dictionary contains the following keys:

Key	Type	Description
max_input_tokens	int	Maximum number of input tokens the model can accept.
max_output_tokens	int	Maximum number of output tokens the model can generate.
text_inputs	bool	Whether the model supports text inputs.
image_inputs	bool	Whether the model supports image inputs.
audio_inputs	bool	Whether the model supports audio inputs.
video_inputs	bool	Whether the model supports video inputs.
text_outputs	bool	Whether the model supports text outputs.
image_outputs	bool	Whether the model supports image outputs.
audio_outputs	bool	Whether the model supports audio outputs.
video_outputs	bool	Whether the model supports video outputs.
reasoning_output	bool	Whether the model supports reasoning output.
tool_calling	bool	Whether the model supports tool/function calling.

Usage Examples

Inspecting Model Capabilities

from langchain_xai.data._profiles import _PROFILES

# Check grok-4 capabilities
grok4_profile = _PROFILES["grok-4"]
print(f"Max input tokens: {grok4_profile['max_input_tokens']}")
print(f"Max output tokens: {grok4_profile['max_output_tokens']}")
print(f"Supports reasoning: {grok4_profile['reasoning_output']}")
print(f"Supports tool calling: {grok4_profile['tool_calling']}")