Google Just Released Two Open AI Models That Can Run on Laptops

Google has been making significant moves in the AI space in an effort to catch up with competitors like Microsoft and OpenAI. In recent weeks, Google has announced several new releases aimed at accelerating its product development. This includes the introduction of Gemini Ultra and Gemini Pro 1.5 under the Gemini brand, as well as the unveiling of two new text-based models called Gemma.

Gemini Ultra stands as Google’s most potent AI offering to date, while Gemini Pro 1.5 boasts comparable performance and an expansive context window for text, images, and audio prompts. The newly launched Gemma models, in contrast, are smaller in size but deliver high performance in strictly text-based tasks. Google emphasizes responsible data handling by removing sensitive information from datasets before training and conducting rigorous testing to address unwanted behaviors.

While built and trained using Gemini technology, Gemma models are uniquely released under an open license, allowing developers to customize and fine-tune them. Google is also providing developer tools to ensure application safety and compatibility across various AI frameworks and platforms. This move aligns with the trend of open models in the industry, catering to developers and organizations of all sizes.

In positioning Gemma as a competitor to Meta’s Llama 2 model, Google differentiates its approach by providing access to model weights for customization without open-sourcing the entire model. This strategy allows Google to strike a balance between enabling innovation and controlling model distribution and ownership. The industry is still navigating the debate between closed and open AI ecosystems, with proponents arguing for the benefits of open collaboration and innovation at scale.

As AI advancements continue at a rapid pace, the landscape is evolving with heightened competition and a focus on responsible AI development. Both closed and open models are driving innovation, and it is expected that big companies and open communities will drive further progress in the AI space throughout the year.