---
license: mit
pipeline_tag: text-generation
library_name: transformers
tags:
- reap
- trinity
---

> [!TIP]
> **[Support this work →](https://donate.sybilsolutions.ai)** · [X](https://x.com/0xsero) · [GitHub](https://github.com/0xsero) · [REAP paper](https://arxiv.org/abs/2510.13999) · [Cerebras REAP](https://huggingface.co/collections/cerebras/cerebras-reap)

# Trinity-337B

REAP-pruned the base model.

## At a glance

| | |
|---|---|
| Base model | — |
| Format | BF16 |
| Total params | **337B** |
| Active / token | — |
| Experts / layer | 216 |
| Layers | 60 |
| Hidden size | 3072 |
| Context | 262,144 |
| On-disk size | 675 GB |

## Which variant should I pick?

| Variant | Format | Link |
|---|---|---|
| `Trinity-337B` **(this)** | BF16 | [link](https://huggingface.co/0xSero/Trinity-337B) |
| `Trinity-337B-W4A16` | W4A16 | [link](https://huggingface.co/0xSero/Trinity-337B-W4A16) |
| `Trinity-337B-W4A16-192` | W4A16 | [link](https://huggingface.co/0xSero/Trinity-337B-W4A16-192) |

## License & citation
License inherited from the base model.

```bibtex
@misc{lasby2025reap,
  title  = {REAP the Experts: Why Pruning Prevails for One-Shot MoE Compression},
  author = {Mike Lasby and Ivan Lazarevich and Nish Sinnadurai and Sean Lie and Yani Ioannou and Vithursan Thangarasa},
  year   = {2025}, eprint = {2510.13999}, archivePrefix = {arXiv}
}
```

## Sponsors
Made possible by **NVIDIA · TNG Technology · Lambda · Prime Intellect · Hot Aisle**.