Arcee AI Trinity Large Is Now Available in Puter.js
On this page
Puter.js now supports Trinity Large, Arcee AI's flagship 400B parameter Mixture of Experts model with frontier-class capabilities in math, coding, and scientific reasoning.
What is Trinity Large?
Trinity Large is an open-weight, 400 billion parameter sparse Mixture of Experts (MoE) model from Arcee AI. Despite its massive scale, it activates only 13B parameters per token using 256 experts with 4 active at any time—delivering frontier performance with 2-3x faster inference than comparable models.
The model natively supports a 512k token context window and was trained on 17 trillion tokens including web, code, math, and reasoning data. Trinity Large outperforms Llama 4 Maverick on benchmarks like MMLU (87.2 vs 85.5) and AIME 2025 (24.0 vs 19.3).
Trinity Large excels at creative writing, storytelling, real-time voice assistance, and complex agent workflows. It supports 14 non-English languages and integrates well with coding agent frameworks.
Examples
Code Generation
puter.ai.chat("Implement a lock-free concurrent queue in Rust with proper memory ordering",
{ model: "arcee-ai/trinity-large-preview:free" }
)
Mathematical Problem Solving
puter.ai.chat("Solve this: Find all positive integers n such that n^2 + 2n + 2 is divisible by n + 1",
{ model: "arcee-ai/trinity-large-preview:free" }
)
Creative Writing
puter.ai.chat("Write a short story about a time traveler who discovers that the future remembers them as a villain",
{ model: "arcee-ai/trinity-large-preview:free", stream: true }
)
Get Started Now
Just add one script tag to your HTML:
<script src="https://js.puter.com/v2/"></script>
No API keys or infrastructure setup needed. Start building with Trinity Large immediately.
Learn more:
Free, Serverless AI and Cloud
Start creating powerful web applications with Puter.js in seconds!
Get Started Now