Nvidia has released a new base model for gaming agents. NitroGen is a vision-action model that has been trained on 40,000 hours of gameplay videos from more than 1,000 games.

The AI research organization METR has published new test results for Claude Opus 4.5. Anthropic's model achieves a so-called 50 percent time horizon of around 4 hours and 49 minutes.

Alibaba's AI unit Qwen has released a new image editing model. Qwen-Image-Layered splits images into several individual layers with transparent backgrounds (RGBA layers).

OpenAI now lets users customize how ChatGPT communicates.

Few areas have seen as much AI progress this year as video generation. Chinese company Kuaishou is closing out the year with new features for Kling 2.6: a technically impressive update that shows just how tight the AI race has become, both internationally and between major players and lesser-known competitors.