Can ChatGPT generate videos? No, OpenAI shut down Sora. Gemini, Grok and Meta AI still can

Why can't ChatGPT make videos anymore?

OpenAI closed the only video product it had, and it named nothing to switch to. On its API deprecations page, the Videos API, sora-2 and sora-2-pro each carry a dash in the column for a recommended replacement. The consumer app went first. Developers lost the API five months later.

OpenAI's Sora discontinuation page says it in one line: "The Sora web and app experiences were discontinued on April 26, 2026." The same page tells former users to export their clips at sora.chatgpt.com/sunset. After any final export window, OpenAI says it will permanently delete the data.

Sora was never a button inside the ChatGPT chat window. OpenAI launched it on 9 December 2024 "as a standalone product at Sora.com to ChatGPT Plus and Pro users," per its launch post. Plus included up to 50 videos a month at 480p, or fewer at 720p. Pro got ten times the usage, up to 1080p and 20 seconds a clip. Sora 2 arrived on 30 September 2025 as an invite-only iOS app in the US and Canada, free with limits at first. That app lasted 208 days.

OpenAI announced the end on 24 March 2026. A spokesperson told Engadget that "We've decided to discontinue Sora in the consumer app and API," and said the Sora research team would keep working on world simulation research for robotics. Engadget also cited Appfigures data: new Sora downloads fell 32% in December 2025 against November.

Today's plans don't hide video behind a higher tier. OpenAI's pricing page, read on 12 October 2026, lists image generation on Free (limited), Go at $8, Plus at $20 and Pro from $100, and no video generation row at all. Its "Voice with video" row is the live camera in voice chats, which lets ChatGPT see what you point your phone at. It's input, not output, and Free doesn't get it.

Which AI chatbots can generate videos instead?

Gemini, Grok and Meta AI generate video inside the chat app as of October 2026, and ChatGPT and Claude don't. The table covers consumer plans in the US, compiled from each vendor's own pages on 12 October 2026.

Gemini is the cheapest of the paid routes. Google's video help page says a personal account "must have a Google AI plan," and the Google AI Plus plan at $4.99 a month includes video generation plus 200 Google Flow credits. You can upload an image or a video as a reference, then edit the result in the same chat by asking.

Grok gives you the longest clips. xAI's SuperGrok page promises "HD 720p, 30-second video" at $30 a month. A Gemini Omni Flash clip in Google Flow tops out at 10 seconds. Meta AI is the route without a subscription. Meta's help center describes making AI videos in the Meta AI app and on meta.ai, but warns that some features aren't available everywhere.

Whizi, which publishes this page, sits in the last row. Whizi Video AI runs Lightricks' LTX-2 model and returns about 4.8 seconds of 720p video with a synced sound track, and it only runs on Powerhouse at $49.99 a month. It loses on both counts that matter here. Grok's clip is six times longer, and Gemini's plan costs a tenth as much. Powerhouse makes sense only if you'd pay for the chat side anyway, since the script and the clip then live in one thread next to GPT and Claude. The credits page lists what each video job spends.

For quality rankings across 29 video models, with prices per second, see the best AI video generator comparison. If you used Sora itself and want the closest replacement, the Sora alternatives page prices 11 of them per second.

Can ChatGPT turn a photo into a video?

No. ChatGPT can make a still image, but it can't animate one, on any plan. Three chat assistants can. Gemini Omni takes an uploaded image as a reference and matches its aspect ratio, per Google's help page. Meta lists "animate images" among its Meta AI help articles. At its launch on 4 August 2025, xAI pitched Grok Imagine as turning text or image prompts into short clips with native audio, TechCrunch reported.

The free option is Google Flow. Every Google account gets 50 Flow credits a day, with or without a subscription, according to Google's Flow credits page. An 8 second Gemini Omni Flash clip at 720p costs 12 credits, so the daily grant covers four of them. Google also warns that free video may be unavailable at peak hours, roughly 2 PM to 5 PM UTC.

What can ChatGPT still do for a video project?

ChatGPT still writes everything a video needs except the pixels: the script, the shot list, the voiceover and the prompts you paste into a generator. That last job matters most now. Each Veo or Gemini Omni Flash clip in Google Flow runs 4 to 10 seconds, so a 30 second video is a chain of shots that must match.

Shot list prompt: "Turn this idea into a shot list for a [length] second video for [platform]: [idea]. Split it into shots of 8 seconds or less. For each shot, give what the camera sees, how it moves, the lighting, and any dialogue or sound. Then write one paragraph prompt per shot that I can paste into a video generator. Repeat the same character and setting description word for word in every prompt, so the clips match."

The repeated description is the part that does the work. Video models don't remember the last clip, so a character described as "a woman in a red raincoat" in shot one and "the woman" in shot two can come back as two people.

Apps inside ChatGPT are the other route. Invideo says its ChatGPT app, called with @invideo, turns a chat into a draft video with scenes, voiceover and music, per its own blog on 27 February 2026. That's the vendor's claim, and invideo renders the video, not OpenAI. ChatGPT will also write animation code (Python with Manim, or Remotion for React) that you run on your own computer.