Why ChatGPT no longer makes videos
ChatGPT made video through Sora, OpenAI's video model. OpenAI announced in March 2026 that it would close Sora: the Sora app and web experience closed on April 26, 2026, and the Sora API was scheduled to end on September 24, 2026. Since then, ChatGPT answers a video request with text or still images.
So when you ask ChatGPT for the bicep heart video, the most it can return is a picture of you in the final pose. That can be enough for a profile photo or a thumbnail, but it will not show the flex and the heart closing.
How to get the moving bicep heart clip instead
Use a photo-to-video tool. On Bicep Heart AI you upload one waist-up photo, pick 480p or 720p, and get a vertical MP4 of about 7 seconds in which your arm comes up into the flex, your other hand reaches across and the heart closes and holds. You do not write a prompt; the pose and timing are built in.
Or film it yourself for free. Prop your phone at chest height, frame yourself from the waist up, flex one arm, curl the other hand high on the bicep and hold for two seconds.
Can I turn ChatGPT's still picture into a video?
You can upload it to a photo-to-video tool, but there is little to gain. A photo that already shows the finished pose gives the AI nothing to animate from: the arm is already up. An ordinary photo with your arms relaxed works better, because the generator moves you from that stance into the flex and the heart.
What about Gemini and other assistants?
Some assistants have their own video models; Google's Gemini app, for example, offers short video generation on some plans. They are general tools: you write the prompt, and the result depends on how well the model handles hands, which the bicep heart relies on. Check the assistant's current limits before you count on it.
