Skip to content

SLM_Survey

A survey and measurement project on small language models

Visit official site
Category AI models

What SLM_Survey is

The project is described as a survey of small language models that combines review with measurement. It covers decoder-only Transformer models between 100M and 5B parameters.

What you can do with it

  • Compare data across small language models
  • Pick a model for a given application
  • Use the findings as teaching material

Getting started

  • Visit the GitHub page for project background
  • Read the README for usage guidance
  • Browse Issues and Discussions for open questions

Pros & cons

✓ What we like

  • Runtime cost is measured through latency and memory
  • Architecture, datasets and training algorithms are analysed

! What to watch out for

  • The scope is limited to decoder-only Transformer models
  • The surveyed set is stated as 59 open SLMs

FAQ

How many models are surveyed?

The material states 59 open small language models.

What is measured?

Inference latency and memory footprint are benchmarked.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review