Can Claude generate images? No, unlike six rival chatbots, but it draws charts

What can Claude make instead of an image?

Claude makes charts, diagrams, data plots and, on Team and Enterprise, short animated MP4 explainers built from code. It won't paint a picture. The FAQ in Anthropic's vision documentation rules out every route to one: Claude "cannot generate, produce, edit, manipulate, or create images." The table sorts each visual job by what Claude does with it, from Anthropic's docs and help center as of October 2026.

The chart route covers more than people expect. VKTR's launch report describes an interactive compound interest curve, built right in the reply. What it can't give you is the thing most people mean by an image: a photo, a painting, a product shot. Anthropic's help center article, updated 16 March 2026, says the same in plainer words.

Which AI chatbots can generate images, and is Claude the only one that cannot?

Claude is the only one of seven paid AI chat assistants whose own app makes no images, as of October 2026. ChatGPT, Gemini, Microsoft Copilot, Grok, Perplexity and Mistral's Le Chat all draw inside the chat, and four of them document it on a free account. Microsoft got there first, on 21 March 2023, when Bing Image Creator put an OpenAI image model in its chatbot.

The table is compiled from each vendor's own help or pricing page, read on 12 October 2026. ChatGPT's image help page says image creation is on all tiers. Google's Gemini help page caps downloads at 1K without an AI plan, and its usage limits page ticks image generation with Nano Banana 2 for accounts with no plan. Perplexity's help center gives Free a limited number of generations. Mistral's pricing page puts basic image generation and editing on Le Chat Free.

Picking which one to draw with is a separate question. The best AI image generators compares them job by job.

How do you get an image while you work in Claude?

You let Claude write the prompt and hand it to a model that draws. That's the route that costs nothing. Paste this into Claude, fill the brackets, and run what it writes in Gemini or ChatGPT:

Image prompt template: "Write a prompt for [Nano Banana, ChatGPT Images or Flux] that produces [the subject] for [a blog header, a product shot, a slide background]. Cover the subject and what it's doing, the setting, the camera angle, the light, the color palette, the style and the aspect ratio. If text must appear in the image, spell it exactly and put it in quotes. Keep it under 80 words. Then give me two variations, one more literal and one more stylized."

Upload the picture back to Claude and ask what to fix. It reads images well, so the review loop works even though the drawing happens elsewhere. Free image quotas are small, and the Gemini image limit explains what a free Google account gets per day.

Three other routes work. Canva's AI connector links Claude to a Canva account, on any Canva plan including Free, and generates and edits Canva designs from the chat. For a flat icon or a diagram, ask Claude for SVG: it writes the code and your browser draws it. And Claude Design, on every plan including Free since 8 October 2026, lays out posters and social posts around images you upload.

Whizi, which publishes this page, puts both halves in one chat. On Pro ($29.99 a month as of October 2026) you can ask Claude Sonnet or Claude Opus for the prompt, then switch the picker to Whizi Image, Nano Banana, Flux or Stable Diffusion in the same thread, with 100 images a month. The pictures still come from those models, not from Claude: Whizi Image and Nano Banana both run a lite version of Google's Nano Banana. Starter gets no images at all. Generate images in a chat shows the steps.

Can Claude read or edit images you upload?

Claude reads images and edits none. On claude.ai you can attach up to 20 images per message, each up to 10 MB and up to 8000 by 8000 pixels, and the API takes up to 600 per request on most models, per Anthropic's vision docs as of October 2026. Claude has read photos, charts and technical diagrams since the Claude 3 launch on 4 March 2024.

Two limits catch people out. Claude can't tell whether a picture is AI generated, and Anthropic's docs say not to rely on it for that. It also won't name the people in a photo. For an actual edit (swap a background, remove an object) you need a model with an edit mode. ChatGPT's image help page lists precise editing, and Le Chat Free includes basic editing.

Will Claude get image generation?

Anthropic hasn't said it will. Claude's release notes through 7 October 2026 add custom visuals on 12 March 2026, Claude Design on 17 April 2026, and designs, decks and documents in any conversation on 16 September 2026. None of them adds a model that draws pictures. The newest visual feature points the same way. Anthropic's 8 October 2026 launch post says Claude Motion "doesn't use a video generation model": Claude writes code that moves the text, charts and images you already have. Claude's pricing page lists no image generation on Free, Pro or Max.

So plan on two tools. Claude does the words, the layout and the charts. A second model does the pixels, and Gemini and ChatGPT both draw on a free account.