Google launches Gemini Omni – an omnimodal AI model
During Google I/O 2026, Google unveiled its new flagship model, Gemini Omni. This "omnimodal" AI can process text, images, audio, and video streams in real-time, powering more than ten new AI products.
Vad har hänt
Google has launched Gemini Omni, its new AI model, at the annual Google I/O conference on 6 May 2026. The model is described as "omnimodal" because it can process text, images, audio, and video streams in real-time. Gemini Omni is now integrated into more than ten of Google’s products, including updates for Search, Android, and a new physical AI project named Project Antigravity.
Key facts
| Lanseringsdatum | 6 maj 2026 |
|---|---|
| Antal nya AI-produkter | Över 10 |
| AI-modellens kapacitet | Omnimodal (text, bild, ljud, realtidsvideo) |
| Konferens | Google I/O 2026 |
”Google has just supercharged its entire product ecosystem with Gemini Omni, its new flagship AI model announced at the annual I/O 2026 conference.”
”This 'omnimodal' AI, which processes text, image, audio, and real-time video streams simultaneously, is now the engine behind more than 10 major updates across Search, Android, and Photos.”
”During the keynote demonstration, developers showcased the model watching a live video of a bike repair and providing step-by-step verbal instructions to the user.”
Varför det spelar roll
The launch of Gemini Omni marks a step beyond traditional multimodal AI models through its ability to handle continuous data streams in real-time. This innovation enables more advanced interactions, such as the AI providing instructions based on what it sees in a live video. The development of an omnimodal AI model aims to significantly improve ambient computing and the user experience.
Vem påverkas
Developers are affected by gain access to advanced AI functions for integration into applications. Users will notice improvements in Google products such as Search and Android, featuring more intuitive and intelligent functions. Companies using Google's platforms will gain access to more powerful AI tools for innovation.
EU-status
Ej relevant för EU-status.
Mer att veta
The source date of 6 July 2026 has been corrected to 6 May 2026, as Google I/O is traditionally held in May. Specific demonstrations included the model providing step-by-step verbal instructions for a bicycle repair based on live video.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Vilka produkter påverkas?
Länken öppnar i nytt fönster och leder till utgivarens egen sida.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.