A community credit grant for developers, researchers, and open source builders working with Trinity models. Apply for free inference access on the Arcee API.
Since we released Trinity-Large-Thinking, developers have taken the model further than we expected. Agent loops running dozens of turns, research prototypes pushing sparse MoE reasoning, open source projects built entirely around Trinity's tool-calling capabilities.
That momentum has also shown up in our support queue. We increasingly hear from builders who have clear projects, clear technical plans, and real community impact but limited inference budget.
Today we're launching the Trinity Builders Program: A community credit grant that gives developers, researchers, and open source builders free inference access to Trinity models on our API.
We built Trinity so developers could own their models. Apache 2.0 weights, open architecture, full control. But open weights alone are not enough. Teams still need compute to run experiments, iterate on evals, and ship real systems.
We want to lower that barrier. If you're building something meaningful with Trinity - A research project, an open source tool, a prototype that could become a product. We want to give you the compute to see it through.
We also have a selfish reason. The builders who run our models in real systems are the ones who surface the insights that shape what comes next. Community feedback on Preview directly informed the direction of Trinity-Large-Thinking, our reasoning-optimized variant post-trained with extended chain-of-thought and agentic RL, purpose-built for the multi-step workflows developers were already running. That feedback loop matters to us, and we want to keep it running.
The program covers the Trinity family models.
Accepted applicants receive free API credits for Trinity models, allocated based on your project scope and our available capacity. During the application you'll select a requested usage range:
Finding the right AI solution shouldn’t feel like guesswork. Our AI Purchase Evaluation Scorecard helps you start your search with clarity.
Extending the Trinity ecosystem or demonstrating novel agent architectures
Pushing the boundaries of reasoning models, sparse MoE inference, or agentic workflows
Using Trinity as the backbone and compute cost is the bottleneck
Making it easier for others to build on top of Trinity
We won't be able to provide credits to everyone; that's a practical reality of running a grant on finite compute. But that doesn't mean you're locked out.
Every Trinity model ships with quantized variants that can run on consumer hardware. If you have a desktop GPU or a small server, you can run the Trinity models locally.
Edge, embedded, and privacy-critical
Cloud deployment
Cloud deployment
Each collection includes the full-precision weights alongside quantized variants (GGUF, W4A16, FP8, NVFP4) suitable for a range of hardware from multi-GPU clusters down to consumer GPUs.
For hardware requirements and setup guides, see Consumer Hardware in our docs.
The last nine months have been about getting Trinity to the frontier. The next phase is about what the community does with it.
We want to be upfront about how this works.
The Trinity Builders Program utilizes the same secure API infrastructure as our commercial tiers, ensuring that no data processed through the program is ever retained or used for model training.
Submitting an application does not guarantee approval. We review requests on a best-effort basis, and allocation depends on compute capacity, token budgets, and other operational priorities. We may provide partial grants, suggest a smaller scope, or decline requests entirely at our discretion.
We reserve the right to modify, pause, or discontinue this program and/or specific grants at any time, for any reason, without prior notice. Awarded credits expire 90 days from allocation and are non-transferable. This is a community initiative, not a contractual commitment. We're running it because we think it's the right thing to do, and we'll keep running it as long as we can.
If you're approved and willing to share feedback or results from your work, that's genuinely appreciated. It helps us prioritize future grants and improve the models but it's not a hard requirement.
