bolt AI and tech · 2 min
Image generation 50% faster, video prices drop 30% this week

flash_on In 20 seconds
OpenAI's new Images 2.6 creates pictures in 6 seconds, adds GIF animation support.
Google's Gemini Omni now makes 15-second videos at $0.07 per second—down from 1.1.
Google's older Gemini models will reroute to newer ones December 1 automatically.
Anthropic now watermarks generated text on three target models for detection.
This week's AI news distills into one shift: generation tools have become so fast and cheap that model swapping, data management, and transparency are now ordinary tasks. Based on a report by Masahiko Studio analyzing official announcements from OpenAI, Google, Anthropic, and Meta, creation speed jumped, prices fell, and old systems began automatic retirements.
The changes touch every role. Image-makers now draft in bulk using speed mode, then polish winners. Video creators pay less per second and iterate through conversation. Developers must audit old model strings before December 1 or face silent rerouting. ChatGPT's finance tools now talk to your bank. The era of "one model forever" is ending.

help What is not known yet
- How long the automatic migration grace period lasts after December 1 for API users who don't manually update
- Exact pricing and rollout timeline for ChatGPT's new financial services by region
- Whether Meta's Spaces will integrate third-party AI tools beyond Meta's own models
forum Frequently asked questions
If I use old Gemini models and do nothing, what happens on December 1?expand_more
Your API calls auto-reroute to newer models at the new model's rate. Output behavior and pricing change without warning unless you update manually.
Can I still create 10-second videos with the new Gemini model, or must I jump to 15?expand_more
You choose duration in 1-second increments between 3 and 15 seconds. Stick with 10 if you want.
Does the new watermark Claude adds to text make it impossible to plagiarize?expand_more
There is no public information on this yet.
How much faster is Images 2.6 compared to the prior version in real time?expand_more
It is 50% faster overall. The fastest mode, Flare at low quality, takes as little as 6 seconds per image.
newspaper Read next
link Sources (1)
- note(ノート) — 7 AI News Highlights from November 2–8, 2026: In an Era of Faster, Cheaper Creation, Migration and Management Become Practical Tasks|マサヒコ|個
How we work: only facts that appear in the cited sources; quotes checked word for word; no opinion and no predictions. Written with the help of artificial intelligence and reviewed by automatic rules.


