Google Translate turns 20, adds pronunciation practice and shares new usage figures

On April 28, 2026, Google marked the 20th anniversary of Translate by launching a new pronunciation practice feature on Android. The anniversary post was less about nostalgia than about a shift in the service's role: from a standalone translator into a layer that understands speech, text and images across the Google ecosystem.

From experiment to platform

The product's history mirrors the evolution of Google's whole AI stack. In 2006, Translate began as a Google Research experiment built on statistical machine learning. The task then came down to finding the most probable phrase matches across huge text corpora. In 2016, it became one of the first Google services to move to neural machine translation, and now the company says outright that it relies on Gemini models and new generations of TPUs to make translation contextual and natural rather than literal.

The scale long ago outgrew an ordinary web tool. According to Google, Translate works with nearly 250 languages, covers 95% of the world's population and supports more than 60,000 potential language pairs. More than 1 billion users a month turn to it for help, and about 1 trillion words a month are translated through Translate, Search, Lens and Circle to Search. Google also stresses its support for endangered and Indigenous languages, an important signal that it sees translation not only as a convenience but as infrastructure for access to knowledge.

A bet on live speech

The main new announcement is a pronunciation practice mode in the Translate app for Android. The feature analyzes the user's speech, gives instant AI feedback and helps them rehearse pronunciation before a real conversation. At launch it is available in the US and India for English, Spanish and Hindi.

It is a logical extension of the existing ask and understand buttons, which provide extra context and translation options. Google is moving steadily away from the "type text, get an equivalent" model toward a language assistant that explains, corrects and teaches.

The same direction shows in live translation. Google says Translate can already work as a personal interpreter in headphones, preserving the tone and rhythm of the other speaker, and that the latest audio-to-audio Gemini models enable smoother real-time dialogue. One telling figure: more than a third of Live translate sessions last longer than five minutes. In other words, people use it not to decipher a single phrase but for real conversations: on trips, on family calls, during tours and even in job interviews.

• Pronunciation practice with instant speech feedback
• Live translate in headphones for real-time dialogue
• More natural translation of idioms, slang and local context
• Offline language packs for trips without internet
• Camera translation in Lens and through Circle to Search on Android

Translation becomes background

The anniversary post shows where Translate lives in 2026. About a third of mobile app users use it not only while traveling but also to learn a new language. Almost half of those who open the Practice feature weekly choose conversation exercises and scenarios resembling real situations. Offline mode remains an important option: the most downloaded languages worldwide are English, Arabic, Spanish, French, Japanese, German, Hindi, Chinese, Russian and Italian.

Even more interesting is how translation dissolves into other Google products. Lens has long turned menus, signs and road markers into a readable layer over the image. On Android, translation is one of the most popular uses of Circle to Search. In Search, people ask AI Mode to explain Gen Alpha slang, translate text into emoji and even help with American Sign Language. Google also revealed the most common language pairs: English → Spanish leads, but the top also includes English → Indonesian, Portuguese, Arabic and Turkish, as well as English → Hindi, Bengali and Malayalam. This is no longer a tourist aid but an everyday interface to the global internet.

"Thank you" remains the most translated phrase in Google Translate.

Alongside it among popular queries are "How are you?", "I love you", "Hello" and "Please".

The list is telling: even with AI assistants, cameras and headphones, people most often translate not business terms but basic phrases for connecting with others. The technology grows more complex, but the core demand stays very human: to understand, to thank, to greet, to explain oneself.

What this means

Google is gradually rebuilding Translate from a standalone service into a multimodal language layer for its whole ecosystem. If this course works, translation will increasingly stop feeling like a separate action and become an invisible part of conversation, search, learning and navigation. That is bad news for competitors, because they will have to fight not an app but Google's systemic advantage in devices, AI models and distribution.