---
title: Hy-MT2 1.8B on SIE | Superlinked
description: "SIE, an open-source inference engine, serves Hy-MT2 1.8B, an open translation model from Tencent that turns English text into your reader's language."
canonical_url: https://superlinked.com/models/tencent-hy-mt2-1-8b
last_updated: 2026-10-29
---

# Hy-MT2 1.8B

An open translation model from Tencent that turns English text into your reader's language.

Model id: `tencent/Hy-MT2-1.8B` · Endpoint: `generate` · [Hugging Face](https://huggingface.co/tencent/Hy-MT2-1.8B)

Listed under [Translation](/models#translation) on [the model index](/models.md).

| Accuracy | Time per request | Serverless price |
|---|---|---|
| 48.1 (WMT24++, 1,488 segments, chrF++) | – (Not timed one request at a time) | $1.13 in · $5.63 out per 1M tokens |

## On the task pages

- [Translate 84% cheaper than Google Cloud Translation](/translation): 48.1 Match with the human translation · 1,488 segments; $3.15 Cost per 1M characters

## Call it

On SIE Serverless:

```python
from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")

text = "Starting in Android 13, apps that you install need your permission to send notifications. Tap to change this permission for existing apps."
result = client.generate(
    "tencent/Hy-MT2-1.8B",
    f"<｜hy_begin▁of▁sentence｜><｜hy_User｜>Translate the following text into Japanese. Note that you should only output the translated result without any additional explanation: {text}<｜hy_Assistant｜>",
    max_new_tokens=256,
    temperature=0,
)
print(result["text"])
```

## Next to the other translation models

| Model | WMT24++, 1,488 segments, chrF++ | Time per request | Serverless price | Size |
|---|---|---|---|---|
| [Hy-MT2 1.8B](/models/tencent-hy-mt2-1-8b.md) `tencent/Hy-MT2-1.8B` | 48.1 | – | $1.13 in · $5.63 out per 1M tokens | 2.0B |
| [madlad400 3b mt](/models/google-madlad400-3b-mt.md) `google/madlad400-3b-mt` | 44.6 | – | $0.999 in · $5.00 out per 1M tokens | 2.9B |

– = not measured / not timed / not listed.

[Both translation models](/models#translation)

## Facts

- Size: 2.0B parameters
- License: Apache-2.0
- Reads: Text, up to 4,096 tokens
- Endpoint: generate

## Evaluation

- Accuracy: Run on SIE against the public dataset named in each figure
- Prices: Serverless list prices
- Sources: [Task page runs and their inputs](https://huggingface.co/datasets/superlinked/sie-task-evidence)