Google launched Gemini Nano on the Pixel 9 series equipped with TensorG4 chipset. It helps smartphones perform artificial intelligence tasks on the device without connecting to a server over the Internet. GeminiNano helps with tasks such as text generation, audio processing, and more. GeminiNano has evolved over time and is now multi-modal, meaning it can understand multiple types of media such as text and images. To take advantage of this capability, MediaTek announced that its flagship processor chip Dimensity9400 will support Gemini Nano with multi-modal capabilities.
It’s worth noting that the feature isn’t limited to the new SoC, but will also be available on other “generative AI-enabled chipsets.”
MediaTek has been working with Google for some time, and the former's 2024 flagship chipset, the MediaTek Dimensity 9400, will be able to handle on-device AI processing. Although no mobile phone equipped with Dimensity9400 has been released yet, it is expected that brands such as Samsung may launch the flagship Galaxy S25 series equipped with this chipset next year.
However, according to MediaTek, it seems that the Dimensity 9400 will not be the company's only chip to support Gemini Nano multi-mode capabilities, and the company is likely to adopt this feature in existing chipsets and future chipsets.
Currently, the Samsung Galaxy Tab S10 series is powered by the MediaTek Dimensity9300+ processor, and it sounds like the chip and the final device will support Gemini Nano with multi-modal capabilities in the future.
The company added: "MediaTek's partnership with Google highlights their shared vision to bring generative AI experiences to the entire Android ecosystem and is a big step forward in running complex AI models on devices. By enabling multi-modal models, users will be able to capture images while receiving detailed descriptions of what they captured."
Currently, there is no information on when MediaTek will release the Dimensity9400 processor and when Gemini Nano will enter MediaTek chips.