Google has made significant updates to its Gemini AI models, including opening access to the 2-million-token context window of the Gemini 1.5 Pro model. Developers now have access to code execution capabilities in the Gemini API and the Gemma 2 model in Google AI Studio. These updates were announced on June 27, with the context window of Gemini 1.5 Pro now available for all developers.
To help developers reduce costs, Google introduced context caching in the Gemini API for tasks that use the same tokens across multiple prompts. Additionally, Google enabled code execution in Gemini 1.5 Pro and Gemini 1.5 Flash, allowing the models to run Python code and learn iteratively until reaching a desired input.
The Gemma 2 model is now available in Google AI Studio for experimentation, and Google is working on offering tuning for Gemini 1.5 Flash to all developers. Text tuning in Gemini 1.5 Flash is ready for red teaming and will be gradually rolled out to all developers. By mid-July, all developers will have access to Gemini 1.5 Flash tuning through the Gemini API and in Google AI Studio.
These updates aim to enhance the capabilities of Google’s AI models and provide developers with more tools for experimentation and optimization.
Source link
Source link: https://www.infoworld.com/article/3715707/google-opens-access-to-2-million-context-window-of-gemini-15-pro.html