Realtime AI videos, new DeepSeek, Nanobanana upgrades, Claude 4.5, realtime TTS, Sora 2 - AI NEWS
The release of real-time open-source models like Wan Alpha and Kani TTS gives indie filmmakers powerful, highly efficient tools that can run locally on consumer hardware.
What this lesson covers
This video covers the latest AI announcements, including Alibaba's Wan Alpha for transparent video generation and Kani TTS for real-time speech. It also reviews updates to Cap4D, Gemini 2.5 Flash, HunyuanImage 3.0, and Claude 4.5.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 00:57 ↗
Wan Alpha enables video generation with native alpha channels and transparency, allowing creators to easily layer elements like fire and water over backgrounds.
- 03:27 ↗
Kani TTS is an ultra-fast open-source text-to-speech model capable of generating 15 seconds of audio in 1 second using an RTX 5080.
- 06:50 ↗
Cap4D allows creators to generate real-time 4D avatars and 3D faces that move over time using multiple source images.
- 08:02 ↗
Gemini 2.5 Flash (Nano Banana) now supports 10 selectable aspect ratios directly within Google's AI Studio.
Workflow outlined in the video
- Visit the Wan Alpha GitHub repository to download and test local transparent video generation workflows.
- Test Kani TTS on Hugging Face to evaluate its multi-language voice synthesis and real-time generation speeds.
- Explore Gemini 2.5 Flash in Google AI Studio to utilize the newly added aspect ratio controls for image generation.
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
100,828 views captured 23 October 2026. ReelStack recommendation score: 38/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 45% performance: 0.98× views vs 9 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: 3 views/day vs 5 channel baseline (8 comparable uploads), with the same view-evidence factor
- 10% engagement: 94/100; likes + 4× comments, smoothed with a 500-view neutral prior