US-based Alphabet Inc., Google’s parent company, has unveiled a new version of its flagship artificial intelligence (AI) model Gemini.
The new Gemini Pro 1.5 model is capable of processing several times more audio, video and text than the GPT-4 model, on which the popular ChatGPT chatbot is based, Wired writes.
In particular, Gemini Pro 1.5 can simultaneously process 1 hour of video, 11 hours of audio, 700 thousand words or 30 thousand lines of program code.
As part of the demonstration, it analyzed the transcript of conversations with the crew of the Apollo 11 spacecraft, presented as a 402-page PDF file, and found several funny moments in it on demand, including the words of the astronauts that the contact was delayed due to a break to eat sandwiches.
Google expects that the new features of the model will allow developers to create new kinds of applications based on it.
The new version of Gemini is already available to them on the AI Studio platform and through Vertex AI’s cloud-based application programming interface (API).
There is no word on when the model will be released to the general public.
Related:
