Weekly AICU Lab+ Study Session! 12/3: NanoBanana for Google Slides & Z-Image-Turbo is Amazing!
Tonight's event [ComfyJapan] December 3, 2025, the 4th session will feature Flux.2 video generation acceleration! And will 'Z-Image-Turbo' become the future champion? Plus, 'NanoBanana Materialization Techniques' for Google Slides, and more practical verification of video generation and AICU-style work efficiency hacks!
Creating Creators! A Recap of the 3rd 'ComfyJapan' Study Session Presented by AICU Lab+!
AICU, 'Creating Creators,' hosts an online study session every Wednesday from 8 PM as 'ComfyJapan' to study video generation using the open-source generative AI tool 'ComfyUI.' We invite VTuber Zassouko Michikusa (Zasuko) (@zasuko_michiksa), a ComfyUI expert, to share the latest tips and tricks in a relaxed talk.
#ComfyJapan 3rd Study Session! #FLUX2 Let's try it out https://t.co/b7UQdj7bIB
— AICU - Creating Creators (@AICUai)
Loading tweet component...
?ref_src=twsrc%5Etfw">November 26, 2025As a summary of the 3rd session on Wednesday, November 26, 2025, we shared topics unique to Japanese users using ComfyUI, including (1) Latest News & InterBEE Report, (2) Ultra-Work Efficiency with 'NanoBanana Pro,' (3) Extremely Heavy! Practical Implementation of the Latest 'Flux.2' Environment.
(1) Latest AI News & InterBEE 2024 Report
At the beginning, we shared a live report from 'InterBEE 2024' held at Makuhari Messe. We introduced AICU's 'Local Video Generation AI Live' exhibit and Zuper ZOO's 'AI Studio Eins,' a dedicated AI animation studio that 'creates emotions by having humans perform.'
We also discussed the latest news regarding copyright (such as the recognition of AI-generated rights infringement and Cloudflare's legal responsibilities). We learned about important legal regulation updates that creators cannot afford to ignore.
(2) Google Slides Creation Becomes Lightning Fast!? 'NanoBanana Pro' Utilization
The demonstration of 'NanoBanana Pro' drew surprised reactions from participants. With a single prompt, 'Visualize the weather and clothing forecast for the next 5 days in Tokyo,' it generated infographics based on accurate weather data.
Furthermore, its dot-picture (Pixel Art) generation ability is also high, instantly creating '80s-style stickers' with quality that could be turned into merchandise as is. Utilization methods that would terrify font craftsmen and logo designers were demonstrated.
Mei Soleil Explains! Nano Banana Pro Official Complete Guide! Explains the innovative functions of Nano Banana Pro. From how to use it in AI Studio to high-definition 4K output, creative j.aicu.ai
In the next study session, we will explore 'NanoBanana Pro' implemented in Google Slides!
(3) Over 200GB of Disk Consumption!? 'Flux.2' Environment Construction and Video Generation Traps
The main dish and 'biggest challenge' this time was the implementation of the latest model, 'Flux.2.'
We attempted to build the environment on Google Colab, but just downloading the model consumed nearly 200GB of disk space, and time melted away just setting it up... Even using an A100 GPU, it took 25 minutes to generate a few seconds of video, revealing the shocking heaviness.
However, the quality of the generated images (such as a composition with three characters lined up) was impressive. We confirmed the high expressiveness unique to Flux, such as 'fusion of photorealism and chibi characters.' This 'human sacrifice' type of verification is one of the real thrills of AICU Lab+.
[Too Fast and Crazy] Alibaba is Serious... 'Z-Image-Turbo' is Born⚡️ Even though the parameters are only 6B Is it true that Flux-class image quality comes out in 8 steps? pic.twitter.com/m3rwaGQS68
— AICU - Creating Creators (@AICUai)
Loading tweet component...
?ref_src=twsrc%5Etfw">November 27, 2025Too Fast!? Alibaba's 'Z-Image-Turbo' is Born!! Alibaba's latest image generation AI 'Z-Image-Turbo' announced. Lightweight and fast, it runs on home GPUs. Technical details and how to use it explained j.aicu.ai
Voices of Participants
'NanoBanana's material creation ability is worth the cost just for this...!'
'Glad I found out how heavy Flux.2 is at Lab+ before trying it on my personal PC (lol).'
'I learn a lot each time, including troubleshooting. I would have definitely given up if I was alone.'
I used Flux2-Dev to generate a 3DCG model in a Colab notebook for note in 88 seconds. It's very fast and comfortable. #AICU #AiCuty #ComfyUI #FLUX2 https://t.co/YUgk3wiKC1 pic.twitter.com/4EfNaTHNiF
— naitochi (@naitochi)
Loading tweet component...
?ref_src=twsrc%5Etfw">November 29, 2025Archives are Available Here
A must-see for those who missed it! You can check the environment construction code for 'Flux.2' and NanoBanana's prompts.
🎁 What is AICU Lab+
An AICU official community where you can learn, share, and present your work on the front lines of generative AI. Linked with magazines, contests, and events, there are also study sessions and benefits limited to members. [ComfyJapan] is the largest ComfyUI research community in Japan.
📺 Next AICU Lab+ Study Session
📅 Next: Wednesday, December 3, 2025, 8:00 PM〜
Theme: 'NanoBanana Materialization Techniques for Google Slides' and 'Z-Image-Turbo' Next time, we will have the highly popular Google Slides creation techniques using 'NanoBanana'! Furthermore, we will also conduct practical verification of the fast-generating 'Z-Image-Turbo.' Let's overcome the busyness of December with AI!
👉 2025/12/3 (Wed) 20:00- AICU Lab+ Study Session 📍 Reservations Here https://j.aicu.ai/ComfyJapan
New Online Program Under Development
For those who have mastered basic operations such as TextToImage and ImageToImage with ComfyUI, we are developing a program where you can acquire practical know-how on environment construction for video production on server GPUs (Google Colab), centering on ComfyUI Wan2.2.
We also offer opportunities to present works and participate in events at the AICU Lab+ community. Deepen your understanding through exercise assignments and knowledge confirmation tests, and aim to improve your skills while interacting with other participants through work submissions and group discussions. Graduates will be issued a certificate, making it ideal for those who want to take a step forward as an AI creator in the video generation era and those who want to prove the basics of video generation using ComfyUI.
Sections (Under Development)
Introduction/Program Overview
Environment Construction with ComfyUI Wan2.2 using Google Colab
Wan2.2 14B TextToVideo / ImageToVideo
Wan2.2 14B Fun Camera Control / Control
Wan2.2 14B Animate Character
Wan2.2 S2V (Speech To Video)
Wan2.2 FLF (First-Last Frame to Video) 14B
Work Submission/Ethics/Compliance
Practical Issues by the Community/Utilization of Certificates
[Session Participation Fee] Spot Participation: 5,500 yen (tax included)
We recommend using the advantageous subscription 'AICU Lab+'. Up to 4 free session participations per month + access to archived videos, plus a PDF version of the monthly 'AICU Magazine' is also provided.
Summary: Why You Should Know ComfyUI and AICU Lab+ Now
As revealed in the 3rd broadcast, the latest AI models (such as Flux.2) are becoming 'larger and more complex' along with higher performance. The time it takes for an individual to download a 200GB file by trial and error and fight errors is now a 'cost.' On the other hand, 'z-image-turbo' is light! fast! high quality! Furthermore, a new model 'Ovis' has also been released.
By participating in the AICU Lab+ study session 'ComfyJapan', you can immediately obtain verified workflows, optimized Colab notebooks, and knowledge of 'what doesn't work.'
In the 4th session on 12/3, we will cover 'Google Slides x AI' that is directly connected to practical work, and the latest high-speed generation technology. If you want to turn AI from a 'playing tool' into a 'work weapon,' please join AICU Lab+!
Hashtags
#ComfyUI #AICULab #Zasuko #MichikusaZassouko #GenerativeAI #AIStudySession #AICUMagazine #ComfyJapan #Flux2 #NanoBanana #ZImageTurbo #GoogleColab
