Arcee Model Overview

Trinity-Large-Thinking is the only Arcee model available via api. See full list of available open models on pricing.

Model	Trinity-Nano (6B)	Trinity-Mini (26B)	Trinity-Large-Thinking (400B)
Strength	Lightweight, ultra-low latency model.	Fast and cost-efficient model for well-defined tasks.	Robust generalist model with strong performance across reasoning, coding, math, and complex task decomposition.
Ideal Deployment	Fully local on consumer GPUs, edge servers, and mobile devices. Tuned for offline operation.	Serve customer-facing apps, agent backends, and high-throughput services in cloud or VPC.	Advanced agents, reasoning systems, and developer tools. Deployed via hosted cloud endpoints or self-hosted in multi-GPU configurations.
Active Parameters	1B per token	3B per token	13B per token
Context Window	128k tokens	128k tokens	512k tokens (hosted at 128k)
Knowledge Cutoff	2024	2024	2024
Speed	⚡⚡⚡⚡⚡ Instant	⚡⚡⚡ Very Fast	⚡⚡⚡ Very Fast
API Model Name	Not Hosted	Not Hosted	trinity-large-thinking