Qualcomm and Meta have announced a new collaboration to bring Llama 3 on Snapdragon-powered devices. It will allow smartphones, VR/AR headsets, vehicles, and other gadgets to run and execute Meta’s in-house AI model locally, enabling on-device AI functionalities. The partnership will also help Snapdragon devices capable of Generative AI, using Llama 3.
Llama 3 on Snapdragon Devices: What it means?
Llama 3 is a custom AI model developed by Meta, similar to GPT models by OpenAI and Gemini by Google. It provides APIs that can be utilized by applications to run AI features. In most cases, these AI models run on cloud servers as they require high computing resources. The device also needs an active internet connection to access those functionalities.
Qualcomm’s latest flagship processors feature dedicated NPUs, like the Hexagon NPU on the Snapdragon 8 Gen 3. It makes the device capable of running and executing AI models on a small scale, which is sufficient for a single user.
The partnership between Meta and Qualcomm opens the doors for on-device AI on Snapdragon devices. This means smartphones or any other compatible product will not require an internet connection for Llama 3-based AI features.
App developers can utilize Llama 3 tools using the Qualcomm AI Hub. Qualcomm maintains a portal that helps developers easily access resources for AI development and research purposes.
Qualcomm has not released the exact list of compatible chipsets supporting Llama 3. We can expect the upcoming Snapdragon 8 Gen 4 processor to have native support for Meta’s AI model. According to Qualcomm’s past launch timelines, the next flagship chip is said to be released around October 2024.
In related news, Meta also announced Meta AI, an AI-powered assistant backed by Llama 3. It is a chatbot similar to ChatGPT, Gemini, and Claude. Meta AI is also accessible to some beta-users of WhatsApp.
