Open-Source SLMs
Coming on the heels of Arcee-Spark – our incredibly performant 7B model – we now bring you Llama-Spark. Built on Llama-3.1-8B, Llama-Spark is a conversational AI that you'd never suspect is just an 8B parameter model.
We're thrilled to announce the latest iteration of Arcee-Spark: Llama-Spark! Built on the foundation of Llama-3.1-8B, this new Spark merges the power of our Tome Dataset with Llama-3.1-8B-Instruct, resulting in a conversational AI that truly punches above its 8B parameter weight class.
Our goal with Spark is simple: to consistently deliver the best-performing conversational AI in the 6-9B parameter range. As new base models become available, we'll continue to update and improve Spark to maintain its leadership position.
We're incredibly grateful for the overwhelming positive response to the original Arcee-Spark, and we're confident you'll love Llama-Spark even more!
A Special Thanks to Prime Intellect
We want to extend a huge thank you to Prime Intellect for being our compute sponsor for this project. We're deeply appreciative of their partnership and commitment to our vision.
.png)
Today we bring you exciting updates on two small language models (SLMs) we've been working on: our first reasoning model, Arcee-Maestro-7B-Preview, and a fast and efficient Mistral-based DeepSeek distillation we call Arcee-Blitz.
Arcee Agent is yet another Arcee model punching above its weight: it's just 7B (initialized from Qwen2-7B) and outranks much larger models. Try it out for function calling and tool use!
Emerging out of intensive research and development, the Arcee team is excited to release our commercial product to contextualize language models in the Arcee platform.
Subscribe to our newsletter
.svg)
