Jason AI
Mini toolsTrinity Nano
New
Text
Gen 7
Overview
Trinity Nano Base is the base 6B (1B active) Trinity Nano checkpoint, pre–fine-tuning, meant to be domain-tuned rather than used directly for chatting, trained on about 10T tokens under Apache-2.0.
Description
Trinity Nano Base is the raw 6B-parameter, 1B-active MoE backbone for Trinity Nano. Trained on roughly 10T tokens (with extra math and code) using large H200 clusters, it is released as a pre–fine-tune checkpoint, so it is not intended as a drop-in chat model. Instead, Arcee positions it for continued training and domain adaptation, with GGUF and other quantized variants available for custom pipelines under the Apache-2.0 license.
Last updated: January 30, 2026
