Can ChatGPT Make Videos? What Works in 2026 After Sora

Scroll To View ↓
Guide8 min readBy Tong Pow
Can ChatGPT Make Videos? What Works in 2026 After Sora

Not by itself. As of November 13, 2026, the ChatGPT plan page lists image generation but no video feature, and OpenAI shut down the Sora 2 video API on September 24, 2026. ChatGPT can still plan a video and direct a connected video tool.

That split matters if you want an ad. ChatGPT is strong at the thinking: the brief, the hook, the script and the scene prompts. The footage has to come from a video service that ChatGPT reaches through a plugin or a custom MCP server.

This guide covers what the OpenAI pages say today, what the retired Sora API offered, and how to get from a chat to an ad-ready video. The sister guide answers the same question for Claude: Can Claude make videos?

Can ChatGPT generate a video?

No ChatGPT plan lists video generation today. The plan comparison on chatgpt.com covers Free, Go, Plus and Pro. It has rows for image generation, voice, deep research and plugins, and no row for creating video (ChatGPT pricing).

The one row with video in the name is "Voice with video", which sits with the voice features.

The overview page tells the same story. It says ChatGPT Work can create and edit "docs, slide decks, spreadsheets, charts, PDFs, images, and other deliverables" (ChatGPT overview). Video is not on that list.

The developer route is closed too. The video guide in the OpenAI API docs now opens with this notice: "The Sora 2 models and Videos API were shut down on September 24, 2026 and are no longer available. No one-to-one replacement API is available" (OpenAI video generation guide).

What ChatGPT and the OpenAI API do for video (checked November 13, 2026)
Video taskStatus todayWhere OpenAI says so
Generate a video clip inside ChatGPTNot listed on the Free, Go, Plus or Pro planChatGPT pricing page
Generate video through the OpenAI APIShut down on September 24, 2026, with no one-to-one replacementVideo generation guide and deprecations page
Generate still imagesLimited on Free; included on Go, Plus and ProChatGPT pricing page
Make a short animated GIFAPI only: 2 to 16 frames with the GPT-Image-2.6 modelsImage generation guide
Use a video tool through a pluginSelect plugins on all plans; some plugins need a paid planPlugins in ChatGPT page
Add a custom MCP serverChatGPT on the web; workspace rules applyAdd custom MCP server guide
Two-column chart of ChatGPT and video as OpenAI lists it: text chats, image generation, voice chats, deep research, plugins and Work files are listed, while video generation, the Sora 2 models and Videos API, and a replacement video API are not listed or closed
What the ChatGPT plan and overview pages name, and what they leave out.

What ChatGPT plans include for video work

Since no plan makes video, the useful question is which plan gives you the pieces around it. The plan comparison shows image generation as limited on Free and included on Go, Plus and Pro. Plus and Pro add "Image generation with Thinking".

Plugins are how ChatGPT reaches outside tools. Discovering and using plugins is marked Yes on all four individual plans. Creating and sharing plugins starts at Go. Developer mode, still labelled beta, is marked Yes only on Plus and Pro (ChatGPT pricing).

Matrix of ChatGPT individual plans Free, Go, Plus and Pro against six rows: image generation, discover and use plugins, create and share plugins, Developer mode beta, Voice with video, and video generation, which is marked not listed on every plan
Video-related rows from the ChatGPT plan comparison. Video generation has no row.

Two things are missing from this section on purpose. The pricing page did not include prices in the HTML that loaded for this check, so no ChatGPT prices appear here. The pages also do not state availability by country for these features. Check the pricing page from your own account before you pick a plan.

What the Sora API did before it closed

Older guides still describe video from OpenAI, so it helps to know what existed. The API changelog shows the Videos API launching on October 6, 2025, with the Sora 2 and Sora 2 Pro models.

On March 12, 2026, an update added clips of up to 20 seconds and 1080p output on Sora 2 Pro (OpenAI API changelog).

The deprecations page then lists the Videos API and every Sora 2 model with a notice date of March 24, 2026, and a shutdown date of September 24, 2026. The replacement column reads "No one-to-one replacement" (OpenAI deprecations).

The consumer side closed earlier. OpenAI's help center says "The Sora web and app experiences were discontinued on April 26, 2026" (OpenAI Help Center).

Timeline of the Sora video API with four dated points: launch on October 6, 2025, longer clips up to 20 seconds on March 12, 2026, the deprecation notice on March 24, 2026, and the shutdown on September 24, 2026
The Sora video API from launch to shutdown, using dates from the OpenAI developer docs.

The retired guide also shows why that API was a weak fit for creator-style ads. Finished clips came back as MP4 files.

The guardrails section says "Real people—including public figures—cannot be generated" and that input images with human faces were rejected. An ad with a founder or a customer on camera was outside those rules.

Where ChatGPT helps with a video ad

ChatGPT handles the work around the render well. The overview page describes drafting "messages, notes, plans, stories, posts, and other writing", and creating images from a prompt to "explore visual directions". For an ad, that maps to four jobs:

  • The brief. Turn a product page and a customer profile into one page on the audience, the offer and the proof.
  • Hooks and scripts. Write several openings for the same offer, then a short script for each.
  • A shot list. Break each script into scenes, with one prompt per scene for a video model.
  • Still images. Make storyboard frames or static ad concepts with image generation.

A brief is only as good as the evidence behind it. Before you ask for scripts, collect ads that already work in your category.

Two free places to look are the Meta Ad Library and the TikTok Ad Library and Creative Center. Paste what you find into the chat, so the scripts follow real patterns.

How to get an ad-ready video from ChatGPT

An ad-ready video is a finished vertical clip with a hook, captions and a call to action, exported as a file you can upload. ChatGPT does not produce that file. A connected tool does, and ChatGPT directs it.

  1. Research what is working and save the best examples.
  2. Write the brief, hooks, script and shot list in ChatGPT.
  3. Connect a video tool as a plugin or as a custom MCP server.
  4. Ask ChatGPT to send the brief to the tool, which renders with its own models and credits.
  5. Review the result, fix captions and claims, then export.
Five-step flow from a ChatGPT chat to an ad-ready video: research what works, write the brief in ChatGPT, connect a video tool as a plugin or custom MCP server, the tool renders with its own models and credits, then review and export
ChatGPT plans and directs; a connected video tool makes the footage.

Using a plugin. OpenAI says you can browse plugins from the ChatGPT sidebar or at chatgpt.com/plugins. You call one from the Tools menu or by typing "@" and the plugin name.

Select plugins are available on all plans, and some require a paid plan. On Enterprise and Edu workspaces, an administrator has to enable plugins first (Plugins in ChatGPT).

Using a custom MCP server. A video service that is not in the directory can still be added by its MCP address. The OpenAI guide gives the path: go to ChatGPT Plugins, select the plus button, then "Add custom MCP server".

Enter a name and the server URL, configure authentication, accept the risk warning and select "Create as a plugin" (Add custom MCP server).

The same guide carries two cautions. It says to connect only to MCP servers you trust. It also says write actions require confirmation by default, so read what ChatGPT is about to send before you approve a paid render.

A good first request names the product, the audience, the format and the number of versions. For example: "Draft three 15-second vertical ads for my budgeting app, each with a different hook. Show me the briefs and the cost before you make anything."

Later, compare the versions with a simple test plan, as in how to A/B test TikTok video ads.

How to connect Ava Studio to ChatGPT

Ava Studio documents a ChatGPT setup. The docs say Ava "works with any app that supports MCP" and that ChatGPT connects with Sign in with Ava, so there is no key to paste (Ava docs: connect your AI app). The steps below follow OpenAI's guide to adding a custom MCP server, with the Ava details from the Ava docs.

  1. In ChatGPT on the web, open Plugins, click +, then "Add custom MCP server".
  2. Name it Ava, paste https://mcp.avastudio.com/mcp and choose OAuth for authentication.
  3. Review the risk warning, then select "Create as a plugin".
  4. Sign in to Ava when asked, check the permissions and click Allow.
  5. Install the plugin in ChatGPT, then mention @Ava in a chat.
Five numbered steps to add Ava Studio to ChatGPT: open Plugins and add a custom MCP server, name it Ava and paste the mcp.avastudio.com address, review the risk warning and create it as a plugin, sign in to Ava, then install the plugin and mention @Ava
Adding Ava Studio to ChatGPT as a custom MCP server, following OpenAI's guide.

Plan support varies. OpenAI's help center says "You don't need to turn on Developer Mode first" in the current Plugins experience. It lists full custom MCP support for Business, Enterprise and Edu, and says Pro users can connect custom servers with read and fetch permissions (OpenAI Help Center). Check your plan before relying on tools that generate media.

The agents page adds that Ava is under review for the ChatGPT plugin directory. Until it is listed, the custom server route above is the way in.

Once connected, ChatGPT can call the tools the API and MCP overview lists. They cover brand setup, research on what is working in short-form video, briefs and shot lists grounded in that research, and campaigns that produce and export videos.

Ava has analyzed 420K+ TikTok ads and posts, and every research result links to the original TikTok. Whatever the agent makes appears in the app, in Ava Campaigns or Ava Studio, for review and editing.

Ava Studio agent chat on a dark screen: the agent offers to extract ingredients, map scenes and prepare prompts, a pasted draft script for an energy drink ad sits below, and a Script Breakdown panel shows an editable text ingredient above the prompt box
A script pasted into the Ava agent becomes editable ingredients before anything is generated.

Connecting does not require a paid Ava plan. Paid operations spend credits from your account. The Free plan is $0, with free campaign research and planning in the Ava app and watermarked 720p exports. From ChatGPT, each research call costs 1 credit and a campaign plan run costs 200. Launch is $59 a month billed annually, or $69 monthly, with 2,000 credits (Ava pricing).

When it is not the right pick: Ava is built for short-form social ads, both paid ad creative and organic posts. It is not a tool for long explainers, slide-style training videos or cinematic films.

For those, a presenter tool or a general video model fits better. The comparison of AI UGC video ad generators covers other options.

When ChatGPT alone is enough

Plenty of video work never needs a render from the chat. ChatGPT alone covers the job when you need a script for someone to film, a storyboard for an editor, or static image ads.

Developers who only need a small loop can use the image API, which returns an animated GIF of 2 to 16 frames (OpenAI image generation guide).

Add a video tool when the deliverable is a finished clip with motion, voice and captions. Pick the tool by the kind of video, since each service uses its own models, limits and pricing. The Claude guide linked above compares twelve video MCP servers.

How we compared

Every OpenAI fact here comes from a page on chatgpt.com or developers.openai.com that was fetched on November 13, 2026. Ava facts come from avastudio.com and docs.avastudio.com, fetched the same day. Nothing in this guide rests on a third-party article, and no ChatGPT plugin from another company was tested.

Two gaps are worth knowing. Most of the OpenAI help center and openai.com blocked automated requests on the day of the check, so the only help article used here is the one on the Sora discontinuation, read on November 14, 2026.

ChatGPT prices and country availability were not in the pages that loaded. Product names and plan limits change often, so confirm them at the source.

Sources

All checked November 13, 2026.

  • chatgpt
  • sora
  • mcp
  • ai video ads
  • ai agents

Frequently Asked Questions

Can ChatGPT make videos for free?

No ChatGPT plan lists video generation, free or paid. The Free plan includes limited image generation and can use select plugins. A connected video service bills its own credits, and some services have a free plan. Ava Studio, for example, has a $0 plan with watermarked 720p exports.

Does ChatGPT Plus include video generation?

The plan comparison on chatgpt.com has no video generation row for Plus or for any other individual plan, as checked on November 13, 2026. Plus does include image generation and plugins. OpenAI lists full custom MCP server support for Business, Enterprise and Edu, so check your plan before adding a video tool.

Is the Sora API still available?

No. The OpenAI API docs say the Sora 2 models and the Videos API were shut down on September 24, 2026, and that no one-to-one replacement API is available. The video guide is kept online for historical reference only.

Can ChatGPT turn images into a video?

Not into a video file. The closest documented feature is in the OpenAI image API, where the GPT-Image-2.6 models can return an animated GIF of 2 to 16 frames. That is a developer feature, and it does not produce a clip with sound.

Can ChatGPT make TikTok video ads?

ChatGPT can write the brief, hooks, script and shot list for a TikTok ad. It cannot render the clip. For a finished ad, connect a video tool through a plugin or a custom MCP server, let that tool render, then review and export the result.

How do I connect a video tool to ChatGPT?

Use a plugin from the ChatGPT plugin directory if the tool is listed. If it is not, go to ChatGPT Plugins on the web, select the plus button, choose Add custom MCP server, enter the server URL and authentication, then create it as a plugin and install it.

Can Claude make videos?

Claude has a different answer from ChatGPT, because it has an animation feature of its own and also works with video MCP servers. The Ava Studio guide called Can Claude Make Videos covers that in detail.

Create your first AI video

Go from a prompt to a polished, viral-ready video with Ava Studio.

Try Ava Studio

Related Articles