Neowin News Feed for: Gemma

Google releases its most powerful open-source AI models yet, that's free to use commercially

Karthik Mudaliar — Thu, 02 Apr 2026 19:16:01 +0000

Google's Gemma 4 is its most powerful open-source model family to date. It beats rivals 20x its size, runs locally on consumer hardware, and comes with a fully unrestricted Apache 2.0 license. Read more...

Google unveils TranslateGemma, a new family of translation models, built on Gemma

David Uzondu — Thu, 15 Jan 2026 21:18:01 +0000

Google shows no signs of slowing its AI advancements, now announcing TranslateGemma, a new set of translation models supporting over 50 languages. Read more...

Google's new AI model discovers a game-changing way to treat cancer

Karthik Mudaliar — Thu, 16 Oct 2025 08:48:01 +0000

Google's new AI model, developed in collaboration with Yale University, has uncovered a revolutionary way to treat cancer. Read more...

Google's new method makes LLMs faster and more powerful, and cheaper too

Paul Hill — Fri, 12 Sep 2025 14:12:01 +0000

Google Research has developed a new method that could make running large language models cheaper and faster. Here's what it has done. Read more...

Google releases Gemma 3n, a new AI model built for mobile devices

David Uzondu — Thu, 26 Jun 2025 19:46:01 +0000

Last month, Google previewed Gemma 3n, an AI model optimized for mobile. Now, it's officially available for developers to test and build with. Read more...

Google unveils Gemma 3n which runs locally on your devices with less memory

Paul Hill — Wed, 21 May 2025 05:02:01 +0000

At Google I/O, the search giant has unveiled its latest small LLM, Gemma 3n. This model is designed to run on your personal devices and uses less memory. Read more...

Google launches DolphinGemma, a new LLM to help understand what dolphins are saying

David Uzondu — Mon, 14 Apr 2025 15:30:01 +0000

Google announced DolphinGemma, an AI model designed to analyze and predict patterns in dolphin communication sounds. Read more...

Google announces Gemma 3, the "most capable model' for a single GPU or TPU

David Uzondu — Wed, 12 Mar 2025 11:02:01 +0000

A year after launching Gemma and Gemma 2, Google has announced Gemma 3, which it claims is the "most capable" model that can run on a single GPU or TPU. Read more...

Google's Gemma 2 delivers powerhouse performance at a fraction of the cost

Pradeep Viswanathan — Thu, 27 Jun 2024 19:28:01 +0000

Gemma 2, particularly the 27B model, delivers class-leading performance comparable to models twice its size. It's designed to run on a single GPU or TPU, significantly reducing deployment costs. Read more...

In a major first, Opera will now allow users to locally run LLMs

Karthik Mudaliar — Wed, 03 Apr 2024 16:44:01 +0000

Opera will allow users using Opera One to run LLMs locally on their computers. The browser supports more than 150 models from 50 families including LLMs from Meta, Google, and more. Read more...

Google reveals another smaller AI model, Gemma, for researchers

John Callaham — Wed, 21 Feb 2024 14:44:01 +0000

Google has just announced yet another kind of language model for generative AI use, Gemma. However, this new model is designed to be used by researchers on their laptops or desktop PCs. Read more...