OpenAI has introduced a new open-weight language model aimed at enhancing advanced reasoning capabilities, marking a significant shift towards greater openness in AI development. This move is designed to engage the global developer community in shaping the model's future, with OpenAI actively seeking feedback and hosting developer sessions to ensure its utility and safety. The model follows in the footsteps of the previously released GPT-2 and responds to the competitive landscape with open-weight models from Mistral, Meta, and xAI. Amazon has launched Nova Act, a model that enables AI agents to perform actions within web browsers, moving beyond simple conversational capabilities. Nova Act, available in research preview at nova.amazon.com, allows developers to create agents that can interact with real-world web interfaces, such as filling forms and navigating dropdowns, using atomic commands and Python interleaving via Playwright. It has achieved over 90% success on tasks like clicking icons and handling calendar pickers on benchmarks like ScreenSpot and GroundUI Web, and is already in use for applications like Alexa+ and async food ordering. Stability AI has expanded its generative media offerings with the launch of Stable Virtual Camera, which converts 2D images into 3D videos, available in research preview via Hugging Face and GitHub. Additionally, Stable Audio 2.0 supports both text-to-audio and audio-to-audio generation on the Stability API. Stability AI has also partnered with WPP to integrate these technologies into media production and branded content. The creator spotlight features Smallfly, which uses SDXL Turbo and Meta’s Quest 3 to transform passthrough footage into AI-powered dreamscapes. PENSIEVE-AI, developed in Singapore, offers a quick and effective method for early dementia detection. By drawing four shapes on a tablet, the AI can identify dementia with 93% accuracy in under five minutes, trained on data from 1,758 seniors. This approach catches 30% more cases than traditional methods and is supported by multilingual audio guides, making it suitable for rapid global screening.