Gemma 3 artificial intelligence bypasses competitors such as Llama, Deepseek and Openai, in systems with a video card, says the company.
Google introduces Gemma 3 – a lighting collection, modern open models created based on both research and technology as Gemini 2.0 models. They are designed for fast work directly on devices, including phones and laptops, and allow developers to create applications with artificial intelligence as needed.
Gemma 3 is provided in four sizes – 1, 4, 12 and 27 billion parameters, giving you the best model for specific needs for hardware and performance. Versions of 4, 12 and 27 billion parameters have support for “vision”, that is, they will be able to work with images and short videos.
There is full support for more than 35 languages and basic training of over 140 languages. There is an expanded context window for 128 thousand tokens, which provides your programs to process and understand the large amount of information.
Additionally it is that -optimize Gemma 3 models to provide maximum performance to graphic processors of any size, from Jetson Nano to the latest Blackwell chips.
The model preceded the Llama-405B, Deepseek-V3 and O3-mini on the initial assessment of human preferences on the LMARENA leaders’ table.
We reminded, earlier reported that China announced the creation of the first autonomous agent of the world AI.
Openai plans to sell access to the science candidate level
News from CORRESPONDENT.NET On the telegram and whatsapp. Subscribe to our channels https://t.me/KorresPondentNet and WhatsApp
Source: korrespondent
I am Ben Stock, a passionate and experienced digital journalist working in the news industry. At the Buna Times, I write articles covering technology developments and related topics. I strive to provide reliable information that my readers can trust. My research skills are top-notch, as well as my ability to craft engaging stories on timely topics with clarity and accuracy.