News

The release of an efficient AI model, Gemma 3, presents a new catalyst for Google. Read what Gemma 3 could mean for AI development and deployment.
While others scale up, Google scaled down — and won. GEMMA 3 is a small model with massive capabilities, showing that ...
Gemma 3 shares tech with Gemini 2.0, supports 35 languages, and offers models with up to 27B parameters. It can be trained through Vertex AI and Google Colab, and it includes a safety image ...
Compared with previous Gemma models, Gemma 3 has expanded from around 8,200 tokens to 128,000. Google also unveiled a new offering for image safety, dubbed ShieldGemma 2, which integrates with ...
In March, Google launched Gemma 3, which the company said can run on a single graphics processing unit or tensor processing unit. In March, it was known that Meta's Llama had hit 1 billion downloads.
At Google I/O Connect India 2025 held in Bengaluru, eight Indian startups showcased cutting-edge applications developed using Google's AI models—Gemini and Gemma. From translation tools and video ...
At Google I/O Connect India 2025, Google also revealed that Invideo is transforming video creation by using Gemini models ...
However, 2B serves Gemma well. Google claims this compact model can run locally without specially designed hardware and outperforms all GPT-3.5 models.
For the same reasons, Google built Gemma to work on a variety of common development platforms. There’s of course a risk that bad actors will use open-source generative AI models to do harm.
The chart illustrates Gemma 2 2B’s superior performance compared to well-known models like GPT-3.5 and Llama 2, challenging the notion that bigger models are always better. (Image: Google) ...