OpenAI Launches ChatGPT Translate: A New Challenger to Google Translate

Without much fanfare, OpenAI has rolled out ChatGPT Translate, a standalone translation tool that directly challenges Google Translate. This new tool is freely accessible without a premium account at chatgpt.com/translate, offering a streamlined interface for translation tasks.

An Interface with an AI-First Approach

The interface of ChatGPT Translate follows the traditional layout of translation tools: two text areas for input and output, automatic source language detection, and support for over 50 languages in both directions. Users can enter text manually, use voice dictation on a mobile browser, or paste content for translation.

What sets ChatGPT Translate apart is its layer of AI-powered customization. After displaying the translation, the tool offers several one-click rephrasing options: making the translation smoother, adopting a formal business tone, simplifying the language for children, or tailoring the text for an academic audience (as shown in the featured image).

By selecting one of these options, users are redirected to the main interface of ChatGPT with a preformed prompt, allowing for further refinement of the translation. This context-, tone-, and audience-centric approach marks a significant departure from traditional translators that focus mainly on linguistic conversion.

Limited Features for Now

Despite its ambitions, ChatGPT Translate currently faces several significant limitations compared to its established competitors. The tool only supports plain text on desktops. Although the page mentions support for image upload translations, this feature is not yet available in the current interface.

The tool also lacks support for documents, handwritten text, websites, or real-time conversations, areas that Google Translate has been developing for years. The number of supported languages is also inconsistent: OpenAI claims over 50 languages on the presentation page, but only 28 were available during our testing, with some duplications like French listed twice, or English missing from the output options while automatic detection works on input.

Regarding voice features, there is no dedicated dictation button in the interface. On Windows, users must use the system shortcut Windows+H, while on mobile browsers, voice entry works but is not natively integrated into ChatGPT Translate.

A Quiet Launch Raises Questions

The lack of official communication from OpenAI and the observed inconsistencies suggest that this might be more of a test version rather than a finalized product. OpenAI has not specified which GPT model is powering the translations, leaving observers to speculate whether it is GPT-5.2 or a model specifically optimized for this task.

This approach contrasts with recent improvements by Google to its translation service. Google recently announced major updates powered by Gemini, including better handling of idiomatic expressions, slang, and local phrases, as well as a beta experience of real-time voice translation through earbuds.