Skip to content

OLMoE

A fully open mixture-of-experts language model

Visit official site
Category AI models

What OLMoE is

The model is described as a fully open mixture-of-experts language model whose data, code and logs are all published.

What you can do with it

  • Run pretraining, fine-tuning and adaptation
  • Use checkpoints and training logs
  • Inspect structure with visual tools

Application scenarios

  • Teams assemble conversational agents
  • Writers produce long form text
  • Translation projects work across languages

Pros & cons

✓ What we like

  • Access is available through the Hugging Face Hub
  • Detailed training and evaluation tools are provided

! What to watch out for

  • Libraries such as transformers and torch must be installed
  • Input data needs conversion before generation

FAQ

How many parameters?

The material states 130 million active and 690 million total parameters.

What is published?

All data, code and logs are described as released.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review