Tags: on-device AI

Enterprise AI firm Cohere launched the Tiny Aya family of open-weight multilingual models, supporting over 70 languages and designed for on‑device use. The base model contains 3.35 billion parameters and runs on everyday hardware without internet connectivity. Regional variants target African, South Asian, and Asia‑Pacific/West‑Asia/European languages. Trained on a single cluster of 64 H100 GPUs, the models are available on HuggingFace, the Cohere platform, Kaggle and Ollama, with accompanying datasets and a forthcoming technical report. Cohere also highlighted strong financial performance and a pending public‑market plan. Weiterlesen

Lightricks introduced an AI video model that runs locally on consumer devices, a rarity in the industry. Built with Nvidia technology and showcased at CES 2026, the model can generate 20‑second clips at 50 frames per second, supports 4K resolution and native audio, and is released as an open‑weight model on HuggingFace. On‑device operation gives creators control over their data, faster generation times of 1‑2 minutes per prompt, and reduced reliance on cloud services, addressing privacy and efficiency concerns for both independent filmmakers and large studios. Weiterlesen

Apple’s new Foundation Models framework, unveiled at WWDC 2025, lets developers embed on‑device AI into iOS 26 applications. The small, local models provide features such as guided generation and tool calling without internet reliance. Early adopters have already integrated the technology into a range of apps—from creative tools for kids and AI‑driven story creators to finance trackers, language learners, task managers, journaling platforms, recipe assistants, and digital signing solutions—enhancing user experiences while keeping data private. Weiterlesen