Ai2 (Allen Institute for AI) models

Ai2 is a non-profit research institute founded by Paul Allen. Its OLMo family is fully open: Ai2 publishes the weights, training data, code and intermediate checkpoints. The current Olmo 3 and 3.1 models come in 7B and 32B sizes with Think and Instruct variants, joined in early 2026 by Olmo Hybrid 7B, which mixes Gated DeltaNet and attention layers.

1 model trackedData checked October 10, 2026

Lineup

Model	Status	Weights	Tier	Context	Input / 1M	Output / 1M	Released
Olmo 3.1 32B ThinkOlmo 3	Current	Open weights	Efficient	64K	—	—	Dec 12, 2025

The most capable tier in the lineup is efficient, led by Olmo 3.1 32B Think. The longest context window is 64K tokens (Olmo 3.1 32B Think).

Frequently asked questions

What models does Ai2 (Allen Institute for AI) offer?

As of October 10, 2026, we track 1 Ai2 (Allen Institute for AI) model: Olmo 3.1 32B Think. Status, prices and context windows for each are in the table on this page.

How do I access Ai2 (Allen Institute for AI) models?

Through Ai2 (Allen Institute for AI)'s developer platform (docs.allenai.org/models/olmo), and for many models through cloud marketplaces and inference providers listed on each model page.

Does Ai2 (Allen Institute for AI) release open-weight models?

OLMo models are released under Apache 2.0, together with their training datasets and code.

Other providers

OpenAI Anthropic Google xAI Amazon (AWS)Meta Perplexity Mistral AI DeepSeek Alibaba (Qwen)Microsoft (Phi)NVIDIA Moonshot AI Z.ai (Zhipu AI)MiniMax IBM

Still choosing a model?

Send us your use case and constraints. We'll come back with a shortlist and an evaluation plan you can run on your own data.

Free and vendor-neutral: no model provider pays for placement.
Your details go only to partners you agree to.