Skip to content

OLMo 2 13B

An open autoregressive model for English benchmarks

Visit official site

What OLMo 2 13B is

The model is described as a Transformer based autoregressive language model developed by Ai2 to support research on language models.

What you can do with it

  • Continue a piece of text such as a story
  • Generate answers within a question answering system
  • Classify and label large sets of documents

Model notes

  • A context length of up to 4096 is supported
  • Fine-tuning options include SFT, DPO and PPO
  • ARC/C, HSwag and WinoG are named benchmarks

Pros & cons

✓ What we like

  • SFT, DPO and PPO fine-tuning options are listed
  • Quantisation is supported for faster inference

! What to watch out for

  • The named focus is English academic benchmarks
  • Integration is described through the Transformers library

FAQ

Who developed it?

Allen Institute for AI, also called Ai2, is named.

What is released with it?

Code, checkpoints, logs and training details are listed.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review