#Google Gemini web app

Last January, Google released a preview of Gemini 2.0 Advanced for paying users, along with the '2.0 Flash Experimental' model. Now, that model has exited the experimental phase and has been officially rolled out to all users on the web and on the Gemini mobile app. Google touts the new Gemini 2.0 Flash as the ideal model for everyday tasks. While Gemini 2.0 Flash is now the default for new conversations, Google confirmed that it will maintain Flash 1.5 and 1.5 Pro for existing conversations for the next few weeks. There is new information that points toward a likely partnership with Google to build its voice assistant off a version of Gemini. Rather than relying strictly on its own hardware, Apple is looking to rely on Google's Gemini AI model running on NVIDIA’s latest Blackwell B200 data center graphics chips. The new system means that when a user triggers a complex prompt that cannot be solved locally on an iPhone, the request will be routed directly to Google Cloud to run on a licensed version of the Gemini model. Originally, rumors suggested Apple might use Google's own custom Tensor Processing Units (TPUs) for these cloud tasks, but the decision has officially flipped to NVIDIA's high-speed Blackwell architecture to ensure maximum performance and lower processing latency. Read more in our articles including "Google Gemini 2.0 Flash now on web, mobile app" and "Apple to revamp Siri with Google Gemini and NVIDIA partnerships".