Today, we are excited to announce that Arcee AI’s Trinity Large-Thinking is now available in Public Preview on DigitalOcean’s Agentic Inference Cloud. This partnership brings advanced reasoning capabilities to developers without the need to manage complex infrastructure or systems.
Trinity Large-Thinking, developed by the San Francisco startup Arcee, offers a family of open-weight models ranging from 4.5B to 400B parameters. In its initial launch, Trinity served over 3.4 trillion tokens on OpenRouter, making it the most-used open weight model in the U.S. This new version, Trinity Large-Thinking, enhances reasoning capabilities, multi-turn tool use, and long-running stability. It ranks #2 on PinchBench for agentic model capability, all at a significantly lower price point than other top-ranking models.
One of the key features of Trinity Large-Thinking is its integration within DigitalOcean’s Agentic Inference Cloud. Developers can now access this advanced reasoning model directly through the DigitalOcean Cloud Console or API, seamlessly alongside their existing compute, data, and services on DigitalOcean.
Reasoning workloads are essential for building agents and complex applications that interpret unstructured data and execute multi-step action sequences. Trinity Large-Thinking offers usable economics, integrated infrastructure, serverless access, and full model control. With this launch, developers can run agents alongside Kubernetes clusters, databases, and storage without the need to stitch together multiple vendors.
This collaboration between DigitalOcean and Arcee marks the next phase of AI infrastructure, where reasoning, data, and compute work together seamlessly. As more workloads shift towards continuous, agent-driven execution, the platform they run on becomes crucial for success.
To learn more about Trinity Large-Thinking and the partnership between DigitalOcean and Arcee, CEO Mark McQuade will be speaking at Deploy on April 28th in San Francisco. You can save your spot to attend live and hear firsthand about the capabilities of Trinity Large-Thinking.
Trinity Large-Thinking is now available in Public Preview on DigitalOcean Serverless Inference. Developers can start running advanced reasoning workloads immediately without the hassle of managing infrastructure or compromising on cost. To get started, simply use the provided curl command to access the Trinity Large-Thinking model and begin running complex reasoning tasks with ease.
This partnership between Arcee AI and DigitalOcean showcases the power of advanced reasoning models in a seamless and integrated AI infrastructure. With Trinity Large-Thinking, developers can unlock new possibilities in AI development and build innovative applications that leverage the power of advanced reasoning capabilities.
For more Information, Refer to this article.
