Trinity models
3 Trinity models from arcee-ai on Hugging Face, from 6.1B to 398.6B parameters, published by arcee-ai in BF16, with contexts from 128K to 256K tokens. The smallest official model, Trinity-Nano-Preview, needs about 13.7 GB of VRAM at its published precision; the cheapest live fit is RTX A4000 at $0.167/hr.
Pick a size
One row per official Trinity size: the VRAM it needs at each precision, the cheapest GPU that holds it today, and the KV cache for a 32K-token context.
| Model | Parameters | Native VRAM | FP8 VRAM | INT4 VRAM | Live GPU fit (native) | Est. $/hr | KV cache at 32K |
|---|---|---|---|---|---|---|---|
| Trinity-Nano-Preview | 6.1B | 13.7 GB | 6.8 GB | 3.4 GB | RTX A4000 | $0.167/hr | 0.52 GB |
| Trinity-Mini | 26.1B | 58.4 GB | 29.2 GB | 14.6 GB | A100 | $1.31/hr | 0.59 GB |
| Trinity-Large-Preview | 398.6B | 891 GB | 446 GB | 223 GB | AMD MI300X × 5 | $13.14/hr | 2.6 GB |
VRAM is weights times a flat 1.2 overhead; the KV cache is a separate per-model figure at 16-bit, shown where the architecture is published. See the methodology.
Official models (3)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| Trinity-Nano-Preview | 6.1B | BF16 | 13.7 GB | RTX A4000 | $0.167/hr |
| Trinity-Mini | 26.1B | BF16 | 58.4 GB | A100 | $1.31/hr |
| Trinity-Large-Preview | 398.6B | BF16 | 891 GB | AMD MI300X × 5 | $13.14/hr |
VRAM is for the precision the model is published in: the weights times a flat 1.2 overhead, with the KV cache not included. See the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.
More from arcee-ai
- arcee-ai (1 models)