---
license: apache-2.0
language:
- en
base_model:
- katanemo/Arch-Router-1.5B
pipeline_tag: text-generation
library_name: transformers
tags:
- text-generation-inference
---

# **Arch-Router-1.5B**

> Arch-Router-1.5B introduces a preference-aligned routing framework that guides model selection by matching queries to user-defined domains (e.g., travel) or action types (e.g., image editing) -- offering a practical mechanism to encode preferences in routing decisions. Specifically, we introduce Arch-Router, a compact 1.5B model that learns to map queries to domain-action preferences for model routing decisions. Experiments on conversational datasets demonstrate that our approach achieves state-of-the-art (SOTA) results in matching queries with human preferences, outperforming top proprietary models.

## Model Files

| File Name | Size | Type | Description |
|-----------|------|------|-------------|
| Arch-Router-1.5B.Q2_K.gguf | 676 MB | Model | Q2_K quantized model (smallest) |
| Arch-Router-1.5B.Q3_K_S.gguf | 761 MB | Model | Q3_K_S quantized model |
| Arch-Router-1.5B.Q3_K_M.gguf | 824 MB | Model | Q3_K_M quantized model |
| Arch-Router-1.5B.Q3_K_L.gguf | 880 MB | Model | Q3_K_L quantized model |
| Arch-Router-1.5B.Q4_K_S.gguf | 940 MB | Model | Q4_K_S quantized model |
| Arch-Router-1.5B.Q4_K_M.gguf | 986 MB | Model | Q4_K_M quantized model |
| Arch-Router-1.5B.Q5_K_S.gguf | 1.1 GB | Model | Q5_K_S quantized model |
| Arch-Router-1.5B.Q5_K_M.gguf | 1.13 GB | Model | Q5_K_M quantized model |
| Arch-Router-1.5B.Q6_K.gguf | 1.27 GB | Model | Q6_K quantized model |
| Arch-Router-1.5B.Q8_0.gguf | 1.65 GB | Model | Q8_0 quantized model |
| Arch-Router-1.5B.BF16.gguf | 3.09 GB | Model | BF16 precision model |
| Arch-Router-1.5B.F16.gguf | 3.09 GB | Model | F16 precision model |
| Arch-Router-1.5B.F32.gguf | 6.18 GB | Model | F32 full precision model (largest) |
| .gitattributes | 2.49 kB | Config | Git LFS configuration |
| config.json | 31 Bytes | Config | Model configuration |
| README.md | 173 Bytes | Documentation | Repository documentation |

## Quants Usage 

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):

![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)