Nvidia has released a new base model for gaming agents. NitroGen is an open vision action model trained on 40,000 hours of gameplay videos from more than 1,000 games. The researchers tapped into a previously overlooked resource: YouTube and Twitch videos with visible controller overlays. Using template matching and a fine-tuned SegFormer model, they extracted player inputs directly from these recordings.
NitroGen builds on Nvidia's GR00T N1.5 robotics model. According to the researchers, it's the first model to demonstrate that robotics foundation models can work as universal agents across virtual environments with different physics engines and visual styles. The model handles various genres—action RPGs, platformers, roguelikes, and more. When dropped into unfamiliar games, it achieves up to 52 percent better success rates than models trained from scratch.
The team, which includes researchers from Nvidia, Stanford, Caltech, and other universities, has made the dataset, model weights, paper, and code publicly available.
Alibaba's AI unit Qwen has released a new image editing model that breaks down photos into separate, editable components. Qwen-Image-Layered splits images into multiple individual layers with transparent backgrounds (RGBA layers), letting users edit each layer independently without affecting the rest of the image.
The model handles straightforward edits like resizing, repositioning, and recoloring individual elements. Users can swap out backgrounds, replace people, modify text, or delete, move, and enlarge objects. Images can be split into either 3 or 8 layers, and the process is repeatable - each layer can be broken down into additional layers as needed. The Qwen team describes this approach as a bridge between standard images and structured, editable representations.
AI research organization METR has released new benchmark results for Claude Opus 4.5.Anthropic's latest model achieved a 50 percent time horizon of roughly 4 hours and 49 minutes—the highest score ever recorded. The time horizon measures how long a task can be while still being solved by an AI model at a given success rate (in this case, 50 percent).
The gap between difficulty levels is big. At the 80 percent success rate, the time horizon drops to just 27 minutes, about the same as older models, so Opus 4.5 mainly shines on longer tasks. The theoretical upper limit of over 20 hours is likely noise from limited test data, METR says.
OpenAI now lets users customize how ChatGPT communicates. The new "Personalization" settings include options for adjusting warmth, enthusiasm, and formatting preferences like headings, lists, and emojis. Each setting can be toggled to "More" or "Less." Users can also pick a base style - like "efficient" for shorter, more direct responses.
OpenAI says these settings only affect the chatbot's tone and style, not its actual capabilities. The company notes that the new options likely work as an extension of the custom instructions feature available in the same settings window.
Google's open standard lets AI agents build user interfaces on the fly
Google’s new A2UI standard gives AI agents the ability to create graphical interfaces on the fly. Instead of just sending text, AIs can now generate forms, buttons, and other UI elements that blend right into any app.
OpenAI is significantly expanding the availability of ChatGPT Go, its budget-friendly subscription tier. Following a launch in India in August, the plan is now available in over 70 additional countries—including markets across Europe and South America—according to an updated support page. In Germany, the service costs 8 euros per month. Beyond extended access to the flagship model, the subscription adds capabilities for image generation, file analysis, and data evaluation, along with a larger context window for handling longer conversations. Users can also organize projects and build their own custom GPTs. However, the plan excludes access to Sora, the API, and older models like GPT-4o.
The broader rollout comes alongside a cost-saving adjustment to how the system handles queries. OpenAI recently removed the automatic model router for users on the free tier and the Go subscription. By default, the system now answers requests using the faster GPT-5.2 Instant. Users must manually switch to more powerful reasoning models when needed, as the automatic routing feature is now exclusive to the higher-priced plans.
