chatgpt-imagesVIDEO
用 GPT Image 2.5 先画分镜再做 AI 视频
作者用 GPT Image 2.5 先出故事板，再交给视频模型生成，避免画面跑偏，附分镜图。
Kirara-Owlbear 🦉
@KiraraOwlbear182026.09.14收录原推发布于 2026.09.14
Kirara-Owlbear 🦉 @KiraraOwlbear18原推发布于 2026.09.14
Raiden Shogun after workout finally hopped on the ai video train since it's everywhere now. the trick for me was using gpt image 2.5 to map out the storyboard first so it doesn't just do its own random thing. Storyboard in the replies minimax H3 on @wavespeed_ai @HiWscreator
x.com/kiraraowlbear18/status/2099348664628158488采集于 2026.09.14 19:37