Trinity Large Thinking
Arcee AI's reasoning-focused Trinity variant. 400B sparse MoE, 13B active per token, designed for complex multi-step reasoning, math, and long-horizon agents. Apache 2.0.
Specifications
SourceRun in the Cloud
This model requires enterprise-grade VRAM. Rent GPUs on RunPod and start generating.
Instant Cloud GPUs
Running out of VRAM? Rent a high-end H100 or RTX 4090 on RunPod and deploy in seconds.
Quantization Estimates
| Format | VRAM Need | Tier |
|---|---|---|
| FP16 | 796.0 GB | Full Precision |
| Q8_0 | 398.0 GB | High |
| Q6_K | 338.3 GB | Excellent |
| Q5_K_M | 278.6 GB | Great |
| Q4_K_M | 199.0 GB | Sweet Spot |
| Q2_K | 119.4 GB | Emergency |
Share this Model
Send these specs directly to your community.
Similar Models
Trinity Nano
6BArcee AI's smallest Trinity model. 6B sparse MoE with 1B active params per token. Fully post-trained for web apps and agent tasks. Apache 2.0.
Trinity Mini
26BArcee AI's 26B sparse MoE with 3B active params per token. Fully post-trained reasoning model for web apps and agentic tasks. Apache 2.0.
Trinity Large
400BArcee AI's frontier 400B sparse MoE with 13B active params per token. One of the largest Apache 2.0 models ever from a US lab. Excels at creative writing, chat, tool use, and long-horizon agentic workflows.