Arcee AI: A US-based Open Intelligence Lab

Three frontier models in six months. All open-weight.

Trinity Models

How We Lead in AI

Models That Matter.

Three releases in six months. Open weights, real benchmarks, no vaporware.

Models That Improve.

Online RL means continuous learning. Deploy once, improve continuously with rapid iteration.

Models That Scale.

Our architectures keep costs low. Frontier performance without frontier pricing.

Our Partners

Nvidia
Intel
ARM
Cisco
Clarifai
Together.AI
Datology.Ai
Prime Intellect
AWS
Microsoft
OpenRouter
Hugging Face

Latest Blogs

Research
•
December 15, 2025

Distilling Kimi Delta Attention into AFM-4.5B (and the Tool We Used to Do It)

Learn how Kimi Delta Attention was distilled into AFM-4.5B using knowledge distillation, long-context training, and Arcee’s open-source DistillKit.

News
•
December 1, 2025

The Trinity Manifesto

Arcee introduces Trinity Mini, a compact MoE model trained end-to-end in the U.S., offering open weights, strong reasoning, and full control for developers.

News
•
October 31, 2025

Mergekit Returns To Its Roots

Effective Friday, October 31, 2025, we are returning Mergekit to the GNU Lesser General Public License v3.

View More