Niantic Spatial is turning 30 billion Pokémon Go player images into a centimeter-precise AI navigation system, now powering Coco Robotics' 1,000 delivery robots across US and European cities where GPS falls short.
Tencent AI Lab launches Penguin-VL, a compact vision-language model that ditches traditional visual encoders in favor of a text-LLM-initialized architecture, delivering stronger fine-grained visual understanding and efficient long-video processing, with two model variants now live on Hugging Face.
MIT researchers unveil a breakthrough AI technique that makes computer vision models both more accurate and explainable by extracting learned concepts and translating them into plain-language descriptions, outperforming existing interpretable models in tasks like bird species identification and skin lesion detection.
Google's Circle to Search and Lens now use a powerful 'fan-out' technique driven by Gemini AI to simultaneously identify and search multiple objects in a single image, delivering one cohesive response in seconds for complex visual queries about outfits, rooms, and more.
Ai2 releases Molmo2, a cutting-edge open-source vision-language model capable of video understanding, object tracking, and pointing across single-image, multi-image, and video tasks, with model sizes ranging from 4B to 8B parameters and fast inference support via Hugging Face Transformers and vLLM.
Apple acquires Israeli AI startup Q.ai for nearly $2 billion in its second-largest deal ever, gaining revolutionary technology that reads facial expressions and micro-movements to enable silent speech communication with Siri through future headphones and glasses.
Google unveils Gemini 3 AI model featuring revolutionary Agentic Vision technology that investigates images like a human detective, zooming into details, executing Python code for annotations, and converting data tables into visual charts with 10% better performance across vision benchmarks.
AI system called AnomalyMatch discovers over 1,400 previously unknown cosmic anomalies in Hubble Space Telescope's 35-year archive, analyzing 100 million images in just 2.5 days to identify merging galaxies, gravitational lenses, and dozens of mysterious objects that defy classification.
Google launches Agentic Vision in Gemini 3 Flash, revolutionizing AI image analysis by enabling the model to actively investigate photos through Python code execution, cropping, and annotation in a Think-Act-Observe loop that delivers 5-10% quality improvements across vision benchmarks.
DeepSeek-AI launches DeepSeek-OCR-2, an open-source visual OCR model featuring groundbreaking Visual Causal Flow technology that mimics human visual processing, supporting dynamic resolution up to 6×768×768 plus 1×1024×1024 image patches with document-to-markdown conversion, PDF processing, and streaming output capabilities through vLLM and Transformers frameworks.