New open Nano Banana, AI plays any video game, new top open source models, long videos: AI NEWS
Open-source tools like FlashPortrait and Nitrogen expand the local toolkit for character animation and pre-visualization without relying purely on closed APIs.
What this lesson covers
This video covers recent open-source AI updates, including Nvidia's Nitrogen video-game playing agent, Alibaba's FlashPortrait for long-form portrait animation, and generative refocusing tools. It details models that improve image editing, 3D scene generation, and video consistency.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 1:11 ↗
Nvidia's Nitrogen is a vision-action foundation model trained on 40,000 hours of gameplay across over 1,000 games to autonomously play video games.
- 3:02 ↗
Alibaba's FlashPortrait generates infinite-length portrait animation videos six times faster than previous methods while maintaining character consistency.
- 6:39 ↗
Generative refocusing tools allow users to take existing photos and adjust focus forward and backward or alter the camera aperture.
- 9:05 ↗
Qwen Image Edit 2511 provides an advanced open-source image editing capability using natural language.
Workflow outlined in the video
- Visit the GitHub repository for Nvidia's Nitrogen to review installation instructions for autonomous gaming agents.
- Check out Alibaba's FlashPortrait GitHub repository to test long-form portrait animation workflows locally.
- Explore the generative refocusing papers and tools to fix depth-of-field and focus issues in source photography.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
How to interpret the numbers
135,980 views captured 23 October 2026. ReelStack recommendation score: 34/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 45% performance: 1.07× views vs 13 other channel videos observed at a similar age. View-evidence factor 99% (views / (views + 1,000))
- 25% freshness: 0/100 with a 14-day half-life
- 20% momentum: 1 views/day vs 5 channel baseline (7 comparable uploads), with the same view-evidence factor
- 10% engagement: 79/100; likes + 4× comments, smoothed with a 500-view neutral prior