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.
World Labs, founded by AI pioneer Dr. Fei-Fei Li, seeks $5 billion valuation amid fierce competition as tech giants pour billions into spatial intelligence technology, believing traditional language models need physical world understanding to achieve human-level artificial intelligence.
Brilliant Labs partners with Liquid AI to power Halo smart glasses with advanced vision-language models that process text and images with millisecond latency, creating personalized knowledge bases and analyzing life context for users.
AbleMouse launches open-source computer control system offering affordable alternatives to expensive assistive technologies, providing tongue, touch, AI vision, and gesture controls for users with ALS, cerebral palsy, and spinal cord injuries across multiple operating systems.
Revolutionary AI tool BMVision slashes kidney cancer detection time by up to 52% while boosting diagnostic accuracy by over 6%, enabling radiologists to catch suspicious lesions earlier during routine CT scans and potentially save more lives through faster, more precise diagnosis.
Thinking Machines Lab launches Tinker platform featuring Kimi K2 Thinking, a trillion-parameter AI model with advanced vision capabilities that outperforms traditional models in image processing and classification tasks with limited data.
Google unveils Gemini 3 Pro, a breakthrough multimodal AI model that delivers state-of-the-art visual reasoning capabilities including complex document processing, pixel-precise spatial understanding, computer screen automation, and high-speed video analysis at 10+ FPS, promising major advances in education, medical imaging, and legal applications.
Dragoneye unveils revolutionary AI Playground that creates custom video detection models in under 5 minutes without any training data, eliminating the need for thousands of hours of video annotation and massive financial investment traditionally required for video AI development.
Anthropic's new Claude Opus 4.5 shatters coding performance records with 80.9% accuracy on industry benchmarks, outperforming OpenAI's GPT-5.1 and Google's Gemini 3 while dramatically cutting costs by 67% and introducing revolutionary 'effort' controls for developers.
Microsoft rolls out new AI agents and Copilot features for Windows including voice activation and smart writing tools, but users revolt demanding the tech giant fix existing bugs and system reliability issues instead of expanding AI capabilities.