Arcee AI released NAC, an open-source harness for managing long-running agent tasks.
News Hub
What actually shipped in agent engineering, pulled from the labs, arXiv and Hacker News.
See who we followArcee AI launched beta access to an open-source models API.
Arcee AI released Genesis-Science-1, a science-focused model.
Trinity is moving to OpenMDW-1.1.
Hermes Agent setup guide using Trinity-Large-Thinking model.
Arcee AI released Trinity-Large-Thinking, a scaled open source frontier agent.
Arcee AI secures strategic investment for enterprise AI platform.
Arcee releases Foundation Model 4.5B.
Arcee AI small language models rank highly on Yupp.ai leaderboards.
Article examines whether running language models on CPU hardware is practical and viable.
Arcee AI released five new open-weight language models.
Arcee AI extended the AFM-4.5B model to support 64,000 token context length.
Arcee AI announced Arcee Foundation Models.
Arcee AI describes tokenizer transplantation as a method to reduce vocabulary barriers between models.
Arcee AI built a retail assistant using small language models on Intel Xeon CPUs.
Arcee AI developed Madeline-s1, a reasoning model.
Arcee AI small language models are now available on Together.ai and OpenRouter platforms.
Llama 4 release received negative reception.
Small language model inference can run efficiently on Arm CPUs.
Arcee AI released Arcee-Maestro-7B-Preview and Arcee-Blitz open-source small language models with reasoning and speed.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.

