Key Takeaways
- **Gemini 1.5 Pro** supports the largest input of any commercially available model, with a capacity of up to 2 million tokens.
- **Trillium**, Google’s 6th generation of TPUs, is the most performant and efficient TPU to date.
- **Google I/O 2024** marked a significant milestone for Google’s AI model series, with breakthrough capabilities and multimodal support.
San Francisco: Google unveiled on Tuesday its latest innovations to its AI model series Gemini at the company’s annual developer conference Google I/O 2024.
The company announced the private preview of a new version of Gemini 1.5 Pro, the company’s current flagship model, that can take in up to 2 million tokens. With that capacity, which doubles the previous maximum amount, the new version of Gemini 1.5 Pro supports the largest input of any commercially available model.
“Today, all of our 2-billion user products use Gemini … More than 1.5 million developers use Gemini models across our tools … We’ve also been bringing Gemini’s breakthrough capabilities across our products, in powerful ways,” Google CEO Sundar Pichai told the conference.
“The power of Gemini — with multimodality, long context and agents — brings us closer to our ultimate goal: making AI helpful for everyone,” he added.
The company has made the improved version of Gemini 1.5 Pro available to all developers globally. 35 languages are now supported in Gemini Advanced, a consumer-facing product that offers 1 million context.
Pichai also announced the company’s 6th generation of TPUs, called Trillium at the conference. “Trillium is our most performant and most efficient TPU to date … We’ll make Trillium available to our Cloud customers in late 2024,” he said.
