MediaTek is working with Meta to integrate the Llama 2 major language model (LLM) into their product. The Taiwanese-based company aims to offer productive use of artificial intelligence on the device that augments the existing feature set.
The Llama 2 model will feature in the next-generation Dimensity processor, which will be released later this year.
Generative AI is everywhere these days, and MediaTek has announced that it’s teaming up with Meta to bring on-device generative AI to the next Dimensity platform. The new processor, which will likely be the Dimensity 9300, will integrate Meta’s Llama 2 major language model, allowing productive AI applications to run entirely on the device without having to go through the cloud.
MediaTek states that “delivering productive artificial intelligence on the device” has several benefits, including seamless performance, greater privacy, better security and reliability, lower latency, the ability to operate in areas with little or no connectivity, and lower operating costs.
MediaTek’s current Dimensity portfolio already includes software that offers generative AI features such as AI-powered Noise Reduction and AI MEMC, and devices that take advantage of the Dimensity 9200, such as the Vivo X90 Pro and Find X6, are highlighting these features.
MediaTek aims to build on this foundation by offering an optimized software stack to run Llama 2, along with an upgraded additional power unit with Transformer backbone acceleration, reduced footprint access, and utilization of DRAM bandwidth to facilitate better productive AI use cases on the device.
The company does not mention the Dimensity 9300 by name, only stating that these features will be included in a next-generation flagship processor that will be released later in the year. We’ll see the Dimensity 9300 launch in November, and MediaTek states that the first phones powered by the hardware will go on sale before the end of the year.
What are you thinking? Please do not forget to share your thoughts with us in the comments.
22150
