Your daily source for AI news, tools reviews, and practical guides to artificial intelligence
Mistral AI introduces Mistral 3, a versatile suite of open-source multilingual and multimodal AI models, ranging from lightweight edge solutions to a powerful large-scale Mixture-of-Experts architecture.
Paris-based startup Mistral AI launches Mistral 3, an innovative family of open-source AI models that span from lightweight edge-friendly versions to a powerful large Mixture-of-Experts architecture, designed to advance multilingual and multimodal AI capabilities.
Zyphra, AMD, and IBM have successfully demonstrated large-scale AI training capabilities using AMD’s GPUs, culminating in the development of ZAYA1, a major Mixture-of-Experts foundation model trained entirely on AMD hardware and networking.
Zyphra, AMD, and IBM have collaborated to demonstrate that AMD’s GPU platform can effectively support large-scale AI model training with the ZAYA1 model, marking a significant step toward diversifying AI infrastructure beyond NVIDIA dominance.
Zyphra, AMD, and IBM collaborated to train ZAYA1, the first significant Mixture-of-Experts AI foundation model built entirely on AMD GPUs, demonstrating a viable alternative to NVIDIA’s dominance in large-scale AI training.
Zyphra, AMD, and IBM have successfully trained ZAYA1, a major Mixture-of-Experts AI foundation model built entirely on AMD GPUs and networking, demonstrating a viable alternative to NVIDIA for scalable AI training.
Zyphra, AMD, and IBM have collaborated to train ZAYA1, the first major Mixture-of-Experts foundation model built entirely on AMD GPUs and networking, demonstrating a viable alternative to NVIDIA for large-scale AI training.