Google launches new version of AI model, along with Search, Gmail integrations

Photo credit: Google

In its latest I/O conference, Google made it clear that it is in its AI era.

The tech giant has launched a wide range of AI products built on the latest iteration of its large language model Gemini, signaling its dedication to compete in the space. The announcement came a day after ChatGPT developer OpenAI’s launch of its new AI model, GPT-4 omni.

Google launched three versions of Gemini. The 1.5 Flash, which the firm says is the “newest and fastest” model, can summarize talks, caption photos and videos, and pull data from documents.

Meanwhile, the 1.5 Pro can take on more complex instructions, with use cases like crafting a chat agent’s “persona and response style.”

Finally, the Nano is a lightweight version of Gemini, which can process image inputs on top of text.

Gemini 1.5 Pro was built with a context window of 2 million tokens, which means it is processing that amount of text to come up with a response. The full 2 million token window is exclusively available to developers, while the 1 million token window of the 1.5 Pro and 1.5 Flash are available to the public for a fee.

Google also announced AI integrations with its Android operating system – including a homework helper system and a genAI assistant – as well as with its Search, Photos, and Workplace products. The latter allows users of Gmail, Google Drive, Google Docs, and other Workplace tools to use Gemini 1.5 Pro through each app’s side panel.

OpenAI’s GPT-4 omni, meanwhile, can accept input and churn out output in any combination of text, audio, and images. With this model, users can have a conversation with ChatGPT directly.

See also: Tether CEO wants to break Microsoft, AWS, Google’s hold on cloud computing