Google Unveils Major Upgrade to Gemini AI with Radically Expanded Capabilities
-
Google announced Gemini 1.5, the next iteration of its AI model, with upgraded architecture and efficiency.
-
Gemini 1.5 Pro has a 128,000 token context window (vs 32,000 in 1.0 Pro) and Google is testing windows up to 1 million tokens.
-
1.5 Pro can process vastly more data at once - up to 1 hour of video, 11 hours of audio, or codebases with 30,000+ lines.
-
In demos, 1.5 Pro could answer complex questions about 402 pages of Apollo 11 transcripts and analyze silent films.
-
Google is running trials of 1.5 Pro and has a waitlist for developers to access it. Wide release will be free up to 125,000 tokens.