Tag
49 articles
This explainer explores Google's breakthrough 'anything-to-anything' AI model that can transform content across different data types using unified multimodal architectures.
Canadian AI company Cohere has open-sourced its most powerful language model to date, Command A+, under the Apache 2.0 license. This move aims to foster collaboration and accessibility in the AI community.
Google introduces Gemini 3.5, a major upgrade to its AI model family that emphasizes 'frontier intelligence with action,' enabling advanced reasoning and tool integration for real-world applications.
ByteDance's Intelligent Creation Lab has released Lance, an open-source unified multimodal model capable of image and video understanding, generation, and editing in a single framework using just 3 billion parameters.
NVIDIA has released Nemotron-Labs-Diffusion, a tri-mode language model that achieves up to six times the token throughput per forward pass compared to Qwen3-8B, offering enhanced efficiency and versatility in AI text generation.
Researchers at the Allen Institute for AI and UC Berkeley have developed EMO, a mixture-of-experts model that maintains near-full performance using only 12.5% of its experts, making it more practical for memory-constrained settings.
Genesis AI, backed by Vinod Khosla, has unveiled its first full-stack AI model GENE-26.5 and demonstrated robotic hands performing complex tasks. The company aims to accelerate AI-powered robotics adoption across multiple industries.
OpenAI is launching GPT-5.5-Cyber, an advanced cybersecurity model exclusively for trusted 'cyber defenders' rather than the general public. The move reflects a responsible approach to deploying sensitive AI technologies.
Chinese AI firm SenseTime releases new image model optimized for speed and domestic chips amid US restrictions. The move reflects a broader trend of Chinese companies building independent AI ecosystems.
Nvidia releases Nemotron 3 Nano Omni, an open multimodal AI model supporting text, image, video, and audio. The model's training data comes from sources like Qwen, GPT-OSS, Kimi, and DeepSeek OCR.
Chinese AI company DeepSeek has released a preview of its next-generation model V4, claiming it can compete with leading US AI systems. The open-source model marks a major leap in coding capabilities and positions DeepSeek as a global AI rival.
OpenAI introduces GPT-5.5, their smartest model yet, designed for complex tasks like coding, research, and data analysis.