Three image tools come up whenever a client asks a designer what they make pictures with. They are Midjourney, Adobe Firefly and the image maker inside ChatGPT, which OpenAI calls . Each is good at a different job. This module shows how to use each one properly, as it works in October 2026, and when to pick which.
It runs in eight steps. This first one sets out what each tool is for, what it costs and who can see what you make. Then come Midjourney set up on its own website, the parameters that change its pictures, its style and character references, and its editor. Firefly on the web follows, with its models, its credits and what Adobe says about commercial use. After that comes ChatGPT for product photographs. The last part runs one brief for No. 4, the face oil in amber glass, through all three tools and compares them honestly.
Midjourney makes pictures with a strong look of their own. Its help pages say that in Standard mode it "automatically adds its own creative touch" to your images, and that Raw mode turns this "auto-pilot" off. The current version is V8.2, the default since 24 July 2026. It lives at midjourney.com, and Discord is now optional. Use it for the look of a campaign, the light, the colour and places that do not exist yet.
Firefly is Adobe's image app at firefly.adobe.com, and the same models sit inside Photoshop. Its menu holds Adobe's own models, Firefly Image 3, 4, 4 Ultra and 5. Next to them are from other companies, among them OpenAI's GPT Image, Google's Gemini and the FLUX models. Adobe says its own Firefly models were trained on licensed content such as Adobe Stock and on public-domain work. That claim is the main reason a client's legal team might ask for it by name.
ChatGPT Images works inside an ordinary chat. You upload a photograph, say what should change, and it draws a new version. OpenAI's help says it can add text, add details or make a background transparent, and since 8 September 2026 it runs on Images 2.5, which OpenAI says keeps the subjects of your reference photos more faithfully than before. It is on every plan, including the free one.
As of October 2026 the money works like this. Midjourney is subscription only, priced in US dollars. Basic is $10 a month, Standard $30, Pro $60 and Mega $120, and paying for a year upfront takes 20 per cent off. Midjourney's subscribe page selects yearly billing by default, so click Monthly Billing before you choose a plan. There is no free trial on the website. Basic gives 200 minutes of a month. Standard and above add , a slower queue with unlimited images.
Firefly counts in generative credits. Free users get a small daily allowance on what Adobe calls a curated set of models. Firefly Standard adds 2,000 a month, Firefly Pro 4,000, Firefly Pro Plus 10,000 and Firefly Premium 50,000, and every paid Firefly plan gives unlimited standard generations, such as Firefly Image 4. Premium credits go on Image 5, Image 4 Ultra and every partner model. Adobe India's plans page draws its rupee prices in by script and could not be read on 9 October 2026, so check it before you pay.
ChatGPT has four personal plans, Free, Go, Plus and Pro. OpenAI's pricing page describes their image limits only in words. Free has "limited image creation", Go "more", Plus "more complex and accurate" and Pro "unlimited and faster", and no plan has a published number of images. The release notes for Images 2.5 say "existing image-generation limits are unchanged". OpenAI's billing help lists rupee billing for India, with UPI accepted on Go and Plus.
Side by side
The three tools side by side, as of October 2026
| Midjourney | Adobe Firefly | ChatGPT Images | |
|---|---|---|---|
| Best at | The look of a campaign before anything is shot: light, colour, mood and places that do not exist. | Changes to real photographs with Adobe's own models, in a file that opens straight in Photoshop. | Changing your own photograph by asking in plain words, one change at a time. |
| Cheapest way in | Basic at $10 a month, billed monthly. No free trial on the website. | Free daily generations on a curated set of models. Firefly Standard adds 2,000 premium credits a month. | The free plan, with limited image creation. Go, Plus and Pro raise the limit. |
| Who sees what you make | Other members, by default, on the Explore page. Stealth mode needs Pro or Mega. | Adobe says it has never trained Firefly on customer content. | OpenAI may use your content to improve its models unless you opt out. |
| Commercial use | You own what you make, but a company earning over $1,000,000 a year must be on Pro or Mega. | Allowed for output from features without a beta label. For partner models, Adobe leaves the call to you. | You own the output, and OpenAI warns that other users may get similar output. |
| Where it lets you down | Exact products. Its Edit model makes a new image from your reference, so the label is redrawn. | The free tier runs on a curated set of models, and every partner model costs premium credits. | Edits can spread. OpenAI's help says highlights are not always precise and changes may reach past your selection. |
Prices, plans and terms from Midjourney's docs, Adobe's generative credits FAQ and AI approach page, and OpenAI's help centre, pricing page and terms, all read on 9 October 2026. Midjourney prices only in US dollars.
What to do it with
- MidjourneySubscription
Generates images from a written description, with the strongest default aesthetic of anything in the category.
Bad at That default aesthetic is the problem as often as it is the draw. It has a house look, it pulls everything towards it, and getting a plain honest photograph of a real product out of it is harder than it should be. Text inside an image is unreliable.
- Adobe FireflyFree tier, paid above it
Adobe's generative models, trained on licensed and public-domain material and wired into Photoshop and Illustrator.
Bad at Raw image quality trails the leaders. What you are buying is the indemnity and the fact that it sits inside files you already have.
- ChatGPTFree tier, paid above it
General assistant with image generation, data analysis and web access in one window.
Bad at Its default voice has become recognisable enough that readers now spot it, which is a brand problem if you ship it unedited.
The words, in plain language
- Fast GPU time
- Midjourney's monthly allowance of quick generations, counted in hours. Basic gets 200 minutes.
- Relax mode
- Midjourney's slower queue, with unlimited images on the Standard plan and above.
- Stealth mode
- The Midjourney setting that hides your pictures from other members. Only Pro and Mega have it.
- Premium credits
- The Firefly credits spent on Adobe's newest model, video and every partner model.
- Partner models
- Image models from other companies, such as OpenAI and Google, offered inside Firefly.
- ChatGPT Images
- OpenAI's name for making and editing pictures inside ChatGPT. Version 2.5 arrived on 8 September 2026.
Why it matters
Pick by the job, never by whichever tool went viral this week. The AI course's first test still sorts everything. Anything a viewer can measure against a real object, such as your product, its label or a face that must recur, comes from a camera or a careful composite. Only the world around it is generated. Within that rule, Midjourney is the strongest at a look, Firefly fits when the file has to end up in Photoshop or a client asks where the training data came from, and ChatGPT is the quickest way to try a change on your own photograph by asking for it.
Know who can see what you make. Midjourney's help says every image you create "can potentially be found" on its Explore page, because the community is open by default, and only Pro and Mega include . That matters the week before a launch. OpenAI's terms say it may use what you put in and get out to improve its services, and that you can opt out of training. Adobe says it has never trained Firefly on customer content. Read these before you upload a product nobody has seen yet.
Know what you are allowed to do with the result. Midjourney's terms say you own what you make, with an exception that catches many studios: a company earning more than $1,000,000 a year must be on Pro or Mega to own its images. OpenAI's terms say you own the output, and add that it "may not be unique", so other users can receive similar pictures. Adobe says output from Firefly features without a beta label can be used commercially. For partner models it leaves the judgement to you, so read that model's own terms first.
People who use these tools every day disagree about how many to pay for. Some keep all three, because each covers a gap in the others and the subscriptions cost less than one day of a photographer. Others keep one and learn it deeply, on the grounds that most bad pictures come from a person who never learnt the controls. Both camps have a point. For a first afternoon, spend almost nothing. Use Firefly's free tier and ChatGPT's free plan, pay Midjourney for one month on Basic, and decide which one earns a place after you have used all three on the same brief.
Elsewhere
- Type
- technique
- Period
- contemporary