Browsing: large language model

Microsoft’s new AI model, BitNet b1.58 2B4T, uses ternary quantization to dramatically reduce memory usage and energy consumption, making it possible to run advanced AI on standard hardware without high-end GPUs.

Alibaba Cloud has launched QwQ-32B, an open-source large language model (LLM) that uses reinforcement learning to achieve performance comparable to models with significantly more parameters.

OpenAI releases its GPT-4.5 model, a large language model designed to improve performance in various areas, despite acknowledging it isn’t a frontier AI model. The release also highlights the ongoing debate about the scalability of AI training.

Hume AI introduces Octave, a new text-to-speech model designed to produce lifelike, emotionally nuanced speech for various content formats, including audiobooks and video games.