Today, June 18th.GoogleWill Gemini 2.5 Pro and Flash launched in a stable version that is fully open to all users, and a cheaper, faster Gemini 2.5 Flash-Lite model, which is currently in preview for "high-throughput tasks such as large-scale categorization or summarization". At the same time, Google brought a key improvement to Gemini -- support for video uploads and analytics.
Users can now upload videos for analysis directly in Gemini for Android, 9to5Google reports. The process is similar to adding an image: just tap the: "+" button and select the video clip from the gallery. Users can ask Gemini to provide an overview of the uploaded video clip, to look for specific parts or objects, etc. In addition to the text reply, Gemini displays the relevant part of the video above the reply. Depending on the length of the video, Gemini's analysis may take some time to complete.
According to 1AI, Gemini's video analytics feature expands on its existing ability to summarize YouTube videos. However, users are currently unable to record videos directly through the Gemini app, and the camera interface still only supports taking and attaching photos to cues.
Gemini's video analytics capabilities set it apart from ChatGPT, which currently only supports analyzing live camera feeds, similar to Gemini Live, and while it can work with images, documents, and PDF files, it does not yet support analytics for video uploads.
Gemini's video uploading and analytics features are currently available on Android and web.