Skip to content

AI Models

Acronym Full Name Category
GPT Generative Pre-trained Transformer Language Model Architecture
SLM Small Language Model Language Model (size class)
HLM Hybrid Language Model Language Model Architecture
LAM Large Action Model Functional Model built on LLMs
LRM Large Reasoning Model Functional Model built on LLMs
VLM Vision-Language Model Multimodal Model (vision + text)
MoE Mixture of Experts Model Architecture Pattern
LCM Latent Consistency Model Generative Image Model (diffusion-family)

GPT - Generative Pre-trained Transformer

  • Dev by OpenAI to power ChatGPT and other generative AI apps
  • Language Language Model (LLM)
    • based on Transformer deep learning architecture

SLM - Small Language Model

  • Smaller version of LLM
    • fewer parameters

HLM - Hybrid Language Model

?

LAM - Large Action Model

  • Agentic AI
    • AI that can perform multiple tasks autonomously
  • Built on top of LLMs
    • Uses LLM to understand and generate human-like text

LRM - Large Reasoning Model

  • AI model designed to perform complex reasoning tasks
    • breaking down problems into smaller steps (resoning traces)

VLM - Vision-Language Model

  • Capable of understanding and processing video, image, text.

MoE - Mixture of Experts

  • Devides AI model into separate sub networks (smaller sub-models ("experts"))

LCM - Latent Consistency Model

  • Generative AI model for creating images
  • Focuses on maintaining consistency in generated images

Sources: