Google produces an updated version of Gimenay AI

Companies are experimenting with this technology which can be used to automate tasks such as programming, summarizing reports or creating marketing campaigns.

Google launched the AI model “Jimenay” last December with three versions, allowing it to be customized to suit the mission to be accomplished, and it can run on all platforms from mobile devices to large-scale data centers.

Jimenay is Google’s response to the strong alliance between Microsoft and Open AI, which some say has been the fastest to take advantage of the current AI boom, including cloud platform customers and developers.

More powerful tools

Now Google is seeking to attract these users to its system using more powerful tools that can be trained “Geminai 1.5” is faster, more efficient and has the ability to process a huge amount of information every time he receives orders to do so according to Venyals, for example, developers can use “Jimenay 1.5 Pro” to query a video of up to an hour’s duration, a 11-hour audio file or a document of more than 700 thousand words, a amount of data says Google is the “longest contextual window” of any large-scale AI model yet.

Jimenai 1.5 can process more data compared to what the latest AI models from Open AI and Anthropic can handle according to Google.

Google showed through a pre-recorded video show prepared for journalists how engineers asked Gimenay 1.5 Pro to accommodate a 402-page document version of the Apollo 11 moon landing and then asked him to find quotes from the document showing “three humorous moments.”

One answer from this AI model suggests that five hours after the “Apollo 11” mission transcript, astronaut Michael Collins told the Mission Control Center: “If we’re late to respond to you, it’s because we’re eating sandwiches.”

In another pre-recorded demo, Google engineers asked Jimenay 1.5 Pro to find a particular scene in the 44-minute film Buster Keaton, based on a rough description of the scene they remember Gimenay successfully found the scene indicating that it was reached about 15 minutes after the video.

Stage Testing

However, Google warned that responses, like all generative models, were not always perfect. Jimenay 1.5 Pro is still fooled and works slowly at times and does not always understand what users mean, forcing them to ask their questions in different ways before the model comes up with the right answer.

Finales said the company was “improving” the performance of Geminai 1.5 “to make it faster, and that it was” still in the process of experimentation and research. ”

The company said that developers can explore “Geminai 1.5 Pro” using Google AI Studio, while some customers of the cloud platform can access the AI model through a special preview on the enterprise platform Vertex AI.

Google also reported on Thursday that it would expand access to its large-scale Gemini 1.0 Ultra model, allowing more global customers on Vertex AI.