Bing AI Image Generator: Credits vs Export-Ready Stills

See how Bing AI image credits work, what a free still includes, and when to finish a keeper in CapCut.

*No credit card required
A drafted graphic beside a finished export
CapCut
CapCut
Nov 13, 2026

A Bing AI image is a Microsoft Designer output, usually reached through Bing or Copilot rather than through a dedicated image app. You type a prompt, the service spends a credit, and you get a still you can download.

That is generation, not search. "Bing images" also names Microsoft's image search engine, and the two behave nothing alike: search returns published photographs with someone else's rights attached, while Designer creates a new file. If you arrived looking for an image to reuse, you are in the wrong workflow—start with the licence, not the prompt.

That credit is the part most guides skip. Generation is metered, retries are metered, and the free allowance is not a permanent quantity—so the way you prompt changes what the session costs. The focus here is spending credits on decisions rather than on experiments a layout tool could have handled.

Table of content
  1. What a Bing AI image actually is
  2. How to use Bing AI images for a still
  3. How credits actually behave
  4. What a free allowance actually gets you
  5. Where the output stops being enough
  6. Microsoft Designer vs a layout editor
  7. Where CapCut finishes a Bing AI image
  8. Entry paths: Designer, Copilot, and Microsoft 365
  9. Conclusion
  10. FAQ

What a Bing AI image actually is

It is a generated still produced by Microsoft's Designer image service, which runs on an external image model and surfaces through Bing and Copilot entry points. The interface changes; the mechanics do not: prompt in, candidates out, credits deducted.

That shape matters because Designer is built for making a graphic, not for finishing a file. It will place a subject on a canvas, add a template, and let you export—but the export is where the limits show: resolution, format, and whether the composition survives a different ratio.

To use Bing images for a real project, treat the session as a drafting budget. Decide the subject and the crop before you type, because every refinement is another credit, and a layout problem is cheaper to solve after generation than during it.

Best for:quick concept frames, social squares, and visual drafts where the composition is still being decided.

Not for:exact typography, print resolution, or a set that has to match a template across ten formats.

How to use Bing AI images for a still

To use Bing AI images for a still, decide the destination ratio, write one prompt naming subject, framing, and light, generate once, then inspect the file at full size before spending another credit. Download the largest size offered and finish the crop and text in a layout tool.

Step
What you do
What to check before spending the next credit
1
Decide the destination ratio first
Will the subject survive that crop?
2
Write one prompt with subject, frame, light
Is anything ambiguous enough to need a retry?
3
Generate once, inspect at full size
Are hands, edges, and any lettering acceptable?
4
Refine one variable, not five
Is the change worth a credit?
5
Download the largest file offered
Is it big enough for the final placement?
6
Finish crop, text, and export in a layout tool
Does the composition still work?
Deciding the destination ratio before generating

The order is the whole method. Most wasted credits go to retries caused by an undecided ratio: the model produces a square, the layout needs a wide frame, and the session restarts.

One habit pays for itself: keep the prompt that produced each keeper. When the next frame has to match, you re-enter a known brief instead of reconstructing one.

How credits actually behave

Designer meters generation with credits, and the allowance is shared across Microsoft's AI surfaces in some plans. Naming and allocation differ between Designer, Copilot, and Microsoft 365, and they change—read the panel in your account rather than a figure quoted in a guide.

The practical consequences:

- Retries are the expensive part. A vague brief burns three credits; a specific one burns one.

- Resolution is not a credit question. You cannot buy your way to print size in a drafting tool; plan the finishing step separately.

- Allowances change. Any number quoted in a guide—including this one—is a snapshot. Check the current panel in your account before planning a batch.

The planning rule that survives every change: generate fewer, better briefs. Two considered renders beat eight exploratory ones, and the difference shows up in both the queue and the quality.

A second rule is less obvious. Credits are consumed by generation, not by thinking, so the cheapest session is the one where the brief was written before the tab opened. Sketch the frame on paper, decide the crop, then type once. Teams that treat the prompt box as a place to think burn allowances on drafts they would have rejected on a napkin.

What a free allowance actually gets you

A free allowance buys renders, not a delivery file. In practice it covers a small number of drafts at screen resolution, a queue you wait in when the service is busy, and whatever the current templates allow on export.

The useful question is therefore not how many images the allowance covers, but how many usable images one session produces. A brief that lands on the first render turns three credits into three assets; a vague brief turns them into three near-duplicates that all need replacing.

Budget the allowance the way you would budget a photoshoot: decide the shot list first, then spend. A session that starts without a crop in mind will always feel too short, no matter what the allowance is.

Where the output stops being enough

Designer is good at producing a composed graphic. It is not a general editor, and four jobs fall outside it:

    1
  1. Recropping for another channel without regenerating the scene.
  2. 2
  3. Setting readable type—headlines and prices belong in editable layers.
  4. 3
  5. Retouching an object edge or a face at full zoom.
  6. 4
  7. Exporting a format or size the current template does not offer.

None of these are failures of the model. They are the boundary between a generation tool and a delivery tool, and recognising it early is what keeps a credit budget under control.

Microsoft Designer vs a layout editor

Dimension
Designer
Layout editor
Core job
Compose an image from a prompt or template
Finish a file for a channel
Cost shape
Credits per generation and queue
Plan-based, per export rules
Text handling
Template text, often raster
Editable text layers
Ratio changes
Re-template or regenerate
Re-crop the same file
Best at
First draft of a graphic
Delivery across formats
Setting a headline as an editable text layer

A useful way to read that table: Designer answers "what should this look like", a layout editor answers "is this file ready". Most teams need both, and the handoff is where the schedule is won or lost.

Where CapCut finishes a Bing AI image

Once the file is downloaded, the work is layout. In CapCut you set the canvas to the destination ratio, crop with the subject inside the safe area, rebuild any headline as editable text, and export at the size the channel needs.

The standard specific to credit-metered generation: spend credits on composition, never on correction. If a fix can be made by cropping or by setting type, it should not cost a render. That single rule is what keeps a Designer session from turning into a slot machine.

For object-level retouching on the downloaded file, CapCut's AI image editor guide covers those moves. For a broader set of generation paths when the free allowance runs out, the free AI image generator guide covers the free-tier breakdown.

Entry paths: Designer, Copilot, and Microsoft 365

The same generation service is reached from three places, and the difference matters for credits rather than for image quality.

Designer is the dedicated surface: a canvas, templates, and an image prompt box. It is the fastest route when the deliverable is a composed graphic.

Copilot puts generation inside a conversation. It is convenient for drafting alongside text, and harder to control, because the canvas is not the main object.

Microsoft 365 exposes the same features inside document apps. Useful when the image has to land next to existing copy; limited when you need a precise crop.

Pick by where the file has to live, not by which entry point sounds newer. And whatever the route, the credit panel in your account is the only reliable statement of what a render costs today.

Inspecting an exported still at full size

Conclusion

Use Bing's Designer for the draft, not the delivery. Decide the ratio first, write one specific brief, spend credits on composition rather than correction, then move the keeper into a layout tool for crop, text, and export. Fewer, better renders are cheaper than any allowance.

FAQ

1. What is a Bing AI image?

It is a still generated through Microsoft Designer, usually reached from Bing or Copilot. This is generation, not Bing image search: it creates a new file from your prompt rather than returning published photos with someone else's rights attached. In the generic sense, an image generator is any system that turns a written prompt into a picture—Designer is Microsoft's implementation of it.

2. How do I use Bing images to generate a picture?

Decide the destination ratio, write one prompt naming subject, framing, and light, generate once, inspect the file at full size, then refine a single variable if needed. Download the largest size offered and finish the crop and text in a layout tool.

3. Is Bing AI image free?

There is a free allowance, and it changes by account, region, and product build. Paid tiers add credits or features. Because the allowance is not fixed, check the current panel before planning a batch of renders.

4. How many credits does one image cost?

Generations are metered per render. The credit system is shared across Microsoft's AI surfaces in some plans, so the name and the allocation you see may differ between Designer, Copilot, and Microsoft 365. Confirm the current figure in your account rather than in a guide.

5. Can I use Bing AI images commercially?

Rights depend on the service terms and your account tier, not on how the image looks. Check the current terms before client work, and keep the prompt and export together so you can show how the file was produced.

6. What is the difference between Designer and a photo editor?

Designer composes a graphic from a prompt or template and hands you a finished-looking file. A photo editor changes the file you already have—crop, retouch, text layers, export settings. The first is a drafting surface; the second is where delivery happens.

7. Why do my Bing images look different from what I asked for?

Usually the brief left a slot unstated—background, camera height, surface—and the model filled it differently than you pictured. Compare two renders of the same prompt: whatever varies is the part you did not specify.

8. Can I get print-resolution files from Bing?

Treat the drafting tool as a screen-resolution source unless the current export offers more. For print, the reliable route is to compose the frame, then finish and export at the size the printer requires, checking the file at 100%.

9. How do I keep a set of Bing images consistent?

Write one brief and change a single slot per image, keeping the rest identical. Then fix the set in the layout: same crop logic, same text position, same export size. Consistency is usually a layout decision rather than a prompt one.

10. When should I stop retrying and change tools?

When the fix you need is a crop, a headline, or an export size—not a new composition. Those are layout jobs, and re-generating for them wastes credits without solving the problem.

Hot and trending