AI News: Google Launches Gemini 1.5 Flash, Updates API and Gemma Line

Google has introduced new features to its Gemini and Gemma product lines, the most important of which is the Gemini 1. 5 Flash model and new features with the Gemini API.

These improvements are meant to be a response to the needs of the developers and enterprise customers who are looking for advanced AI functions for different applications.

Gemini 1.5 Pro and Introduction of Gemini 1.5 Flash

Google has rolled out improvements to its Gemini 1.5 Pro model and launched the new Gemini 1.5 Flash. The updated Gemini 1. 5 Pro now has the longer context window of up to 2 million tokens which is the basis of the more complex and subtle instructions across a variety of tasks such as code generation, logical reasoning, and multimodal interactions.

This improvement is meant to expand the users’ possibilities and to get more thorough analysis.

Introducing Gemini 1.5 Flash ⚡

It’s a lighter-weight model, optimized for tasks where low latency and cost matter most. Starting today, developers can use it with up to 1 million tokens in Google AI Studio and Vertex AI. #GoogleIO pic.twitter.com/I1adecF9UT

— Google (@Google) May 14, 2024

The new Gemini 1.5 Flash is made to fulfill the requirements of high efficiency and low latency for applications which need the rapid response. The Gemini 1.5 Flash is customized for the high-frequency tasks and it is a lighter and faster option to the Pro model, which is more robust, therefore, it is perfect for tasks like data extraction from long documents, summarization, and real-time chat applications.

This model works through a method called “distillation” to make sure it stays at the level of high quality while it is moving at a faster speed.

Expansion of the Gemma Model Family

Besides the updates to the Gemini line, Google has also enlarged its Gemma series of open models. The introduction of Gemma 2 is a major advance in the development of open models aimed at the high performance and efficiency on different computing platforms.

Gemma 2, which is going to be launched in June, has a 27 billion parameter configuration that is designed for both GPUs and TPUs, and it is expected to be twice as good as models that are twice its size.

Besides that, the Gemma family is also happy to introduce PaliGemma a new vision-language model which is optimized for image captioning and visual Q&A tasks. The enlargement of Google is the proof of the fact that Google is committed to making the AI tools which are widely used and useful for the researchers to large enterprises.

New Features and Pricing for the Gemini API

The Gemini API has been developed in response to the user feedback and it now has several new features in it. These are video frame extraction and parallel function calling which enable the developers to execute multiple function calls at the same time. These features are set to be released in June and they are aimed at the improvement of the performance and user experience of the Gemini models.

Besides the new features, Google has also changed the pricing for the Gemini API. The API will be still free of charge in the regions that are eligible through Google AI Studio with the increased rate limits available in the new pay-as-you-go service.

These changes are intended to make advanced AI capabilities more accessible to a wider range of users and to encourage innovation using Google’s AI technology. This development follows OpenAI’s release of ChatGPT 4.0, which has been made available to all users, including those on free plans. Google’s latest updates can be seen as a strategic move to maintain its competitive edge. 

Read Also: Stellar (XLM) Testnet Upgraded To Protocol 21, Will It Impact Price?