MediaTek announces support for Gemini Nano multimodal for select chips

MediaTek is adding Gemini Nano support to its processors, beginning with the soon-to-be-released Dimensity 9400

MediaTek is integrating Gemini Nano support into its processors, starting with the upcoming Dimensity 9400. The company has announced that the Dimensity 9400 will be optimised to support Gemini Nano, featuring multimodal capabilities. The Taiwanese fabless semiconductor company said that this support will extend beyond the new chipset to include other Generative AI-enabled chipsets as well. However, it remains unclear which processors will support this functionality.

“MediaTek’s partnership with Google highlights a shared vision for bringing Generative AI experiences across the Android ecosystem and represents a significant advancement in running complex AI models on-device,” stated MediaTek.

 

Gemini Nano has become available on more devices over the past few months. The multimodal Gemini Nano made its debut on the Pixel 9 series, powered by Google’s Tensor G4 chipset, and is utilised for features such as Pixel Recorder. With the help of multimodal AI, Google’s Generative AI can operate on a smaller scale across devices.

The MediaTek Dimensity 9400 chipset is set to launch in October. The upcoming Dimensity 9400 will feature a new 8th generation Neural Processing Unit (NPU) that offers hardware acceleration for text, image, and speech processing. According to the announcement, it appears that the Dimensity 9400 will not be the only chip with this support; devices running on MediaTek’s existing and future chips will also have compatibility.

MediaTek noted that with the addition of multimodal capabilities, users will be able to take images and receive detailed descriptions of what has been captured.