# TranslateGemma (google/translategemma-4b-it)

> Open translation model by Google DeepMind lightweight enough for on-device use.

TranslateGemma is an open language model variant from Google DeepMind optimized for text translation across dozens of languages. At 4B parameters, it is lightweight enough to be deployed on local hardware and mobile devices using frameworks like Apple MLX.

Website: https://huggingface.co/google/translategemma-4b-it

## What it is for

- Run offline, on-device translation in mobile applications
- Translate text across 50+ languages without sending data to third-party APIs
- Deploy a dedicated local translation model on Apple silicon hardware

## Features

- 4B parameter instruction-tuned model size
- Support for translation across 50+ languages
- Compatible with Hugging Face Transformers and Apple MLX
- Runs locally on edge and consumer hardware like iPhones and Macs

## Pricing

Open source. Free under Gemma license terms

## Details

- Type: AI model
- Category: [AI](https://xiry.app/topics/ai.md)
- Open source: yes
- Platforms: Code library, Self-hosted, iOS, macOS, Linux
- For: Mobile and AI developers building offline or private translation features.
- Alternative to: Google Cloud Translation API, DeepL API, NLLB-200
- Tags: translation, gemma, deepmind, on device, mlx, nlp

## Why it was saved

- [@adrgrondin, 26 January 2026](https://x.com/adrgrondin/status/2015886566216900938): Named in the tweet as the model powering on-device translation on iPhone; relevant for anyone building local AI translation features.

## Related

- [Locally Translate](https://xiry.app/resources/locally-translate): Offline, private AI translation app running directly on your iPhone. (Mobile app, Free)
- [TranslateGemma 4B WebGPU](https://xiry.app/resources/translategemma-4b-webgpu): Private, in-browser machine translation running Google's TranslateGemma model via WebGPU (Web app, Free)
- [HeyGen](https://xiry.app/resources/heygen): AI avatar video generation and translation platform driven by text prompts (Web app, Freemium)
- [Google DeepMind Veo](https://xiry.app/resources/google-deepmind-veo): Generative AI video model producing cinematic footage with native audio (AI model, Freemium)
- [Agentic Video in Gemini](https://xiry.app/resources/agentic-video-in-gemini): Dynamic video inspection for Gemini models that cuts token usage and cost (AI model, Paid)
- [Auris AI](https://xiry.app/resources/auris-ai): AI speech-to-text platform for multilingual transcription, translation, and video subtitling (Web app, Freemium)

Updated 10 October 2026. Page: https://xiry.app/resources/translategemma-google-translategemma-4b-it

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.