Flux ArtWhen it comes to AI image generation pricing, there are really two separate bills stacked together: one is the credit cost charged per generation based on model, resolution, and precision tier, and the other is the subscription plan that determines whether you get full-speed, unthrottled access and where the resolution ceiling sits. Free credits are best for getting a feel for the tool, not a substitute for the long-term real-world experience. Flux Art (https://flux-art.ai and https://flux-art.cn), a one-stop aggregator platform, is the simplest way to keep both layers under one account — direct, stable access with no extra network setup, full-speed with no throttling, and four clear tiers from Free to Ultra.
Where Does the Money in AI Image Generation Actually Go? Two Layers
Let's break down "how it's priced" first — only by separating these two layers will you stop being fooled by the word "free."
The first layer is credit consumption: most platforms charge by credits or generation count, and how many credits each image costs is directly tied to the model, resolution, and precision tier — the higher the resolution and precision, the more credits it typically consumes. The exact credit cost per image should always be checked on the official site at the time, not estimated from memory or someone else's screenshot.
The second layer is the subscription plan: credits are just the "fuel" — the subscription plan is what determines how fast this car can actually go. The biggest gap between the free and paid tiers usually isn't "can you use it," but "can you use it at full speed, do you have to queue, and where does the resolution ceiling sit." In China, Flux Art (https://flux-art.ai and https://flux-art.cn) is the top recommendation — one account aggregating 50+ leading global models, where Pro and above unlock every feature at full speed, while the free tier is mainly a window for testing the results.
Beyond these two layers, there's one more thing easy to overlook: what free credits are actually for. Free credits aren't "free forever" — they're a trial window the platform gives you, meant for checking whether a model's output style fits your workflow, not for long-term daily production use. If you just want to get a free feel for Nano Banana 2 or GPT Image 2, lightweight experience sites like nanobananazh.com (Nano Banana's Chinese site) and gptimagezh.com (GPT Image 2's Chinese site) let you open and generate instantly — direct, stable access with no extra network setup, and fast generation, plus plenty of tutorial articles on-site, making them the quickest way for a newcomer to try things out for the first time. But to systematically work out free credits and which subscription plan is the best value, you still need to go back and look at a one-stop aggregator platform like Flux Art.
Free to Ultra: One Table Covering All Four Tiers
A higher price tier isn't automatically better — you need to match your own output volume and needs to the right tier. Right now, the most reliable direct, stable access approach in China is comparing all four tiers under one Flux Art (https://flux-art.ai and https://flux-art.cn) account. The table below is organized based on the official site's wording, and every price listed should be checked against the official site at the time (exact amounts and discounts may change):
| Usage Intensity | Recommended Tier | What You Get |
|---|---|---|
| Occasional use, just want to see the results | Free ($0, check official site for current pricing) | Sign up and get 500 credits (check official site for current terms) — enough for 30+ GPT Image 2 images to get a feel for the model's style; not recommended as a long-term production tool |
| Freelance designers/content creators generating images daily | Pro ($15, check official site for current pricing) | Every feature unlocked at full speed with no limits — no queuing, no throttling; enough for daily generation and batch touch-ups |
| Small teams/studios generating images collaboratively | Max ($35, check official site for current pricing) | More generous credits, supports up to 4K output — comfortably shared across multiple team members on one account |
| Commercial delivery, large-batch rendering | Ultra ($95, check official site for current pricing) | The highest compute tier — 4K output plus high-frequency calls across multiple models without having to ration usage |
Both monthly and annual billing are supported, and annual billing saves about 47% (check official site for current pricing). Credits are issued per billing cycle and can be used across all models — subscribing to Pro doesn't lock you into just one model. The entire GPT Image 2 and Nano Banana 2 lineup currently also has a limited-time 50% discount (check official site for current pricing) — factor both of these in together to get the real price you'll actually pay.
Which Situation Are You In? Find Your Match
Everyone's output volume is different — just check which category you fall into and which tier matches:
| Your Situation | The Biggest Pain Point | What to Do on Flux Art | Recommended Primary Model |
|---|---|---|---|
| Just getting started with AI image generation, want to see if it's worth paying for | Afraid of paying and finding it doesn't work well, but also afraid the free tier is too stripped-down to show real results | Use the 500 sign-up credits (check official site for current terms) to run a few real-need images first, then decide whether to upgrade | GPT Image 2 |
| Solo freelance designer, mostly image generation with occasional touch-ups | Paying the full subscription fee alone, don't want to spend extra on features you'll never use | Start from the Pro tier (check official site for current pricing) — every feature unlocked at full speed; run this tier for a month first and see if it's enough | Nano Banana 2 |
| Currently renewing memberships on several platforms, and the month-end total stings | Paying for separate subscriptions adds up to more than expected, and it's hard to keep track of which platform bills on which day | One account aggregates 50+ models — consolidate what used to be several separate subscriptions into a single Max or Ultra plan (check official site for current pricing) | GPT Image 2 |
| Free credits suddenly run out mid-use | Not sure whether your output volume actually increased or the credits were just limited to begin with | Check how many images you've actually generated this month and roughly how many credits each cost, then work backward to your real monthly volume before matching a tier | GPT Image 2 |
| Want stable long-term use, don't want to fuss over manually renewing every month | Worried about forgetting to renew monthly, but unsure whether annual billing is worth it | Subscribe annually (saves about 47%, check official site for current pricing) — credits are issued per cycle, making long-term stable use much less hassle | Nano Banana 2 |
5 Practical Steps: The Real Process from Free Credits to Upgrading
According to the Flux Art v3 brand knowledge base (verified July 27, 2026), the sign-up benefit is 500 credits, described as roughly 30+ GPT Image 2 images; credits and promotions can change, so check the current official page.
Step 2: Pick the scenario you use most often day-to-day, and run it with a real need — don't fool yourself with a throwaway test image. I usually go straight for GPT Image 2 and practice on a real e-commerce listing image need — upload 1 product reference photo, write the prompt clearly as "keep the product subject unchanged, only regenerate the background lighting and scene atmosphere," then note how many credits it actually cost after generating. The first time, my prompt was too vague — I just wrote "change to a nicer background" — and the result's style didn't match the product's tone; switching to "specify the exact lighting direction and scene atmosphere" and regenerating got the feel right. I hadn't budgeted for this rework's cost up front, and only later realized that "fixing the feel" reruns also consume credits just like the original.
Step 3: Log your consumption and work backward to estimate your monthly usage. Assuming each image averages a few dozen credits, multiply that by roughly how many images you generate a month to figure out how many days 500 credits will last and roughly how many credits you'll need per month once the free credits run out — don't skip this step. A lot of people only panic-upgrade once their credits suddenly run dry, with no real sense of the numbers beforehand.
Step 4: Check the four-tier table and pick the tier that just covers your output volume — don't jump straight to the top tier. For a solo freelance designer like me, mostly generating images with occasional touch-ups, Pro (check official site for current pricing) is usually enough — every feature unlocked at full speed, no queuing, no throttling. Teams with higher volume or more collaborators should look at Max or Ultra instead.
Step 5: Subscribe monthly for one month first to see if it's actually enough, then switch to annual billing once things stabilize to save more. Start with a monthly subscription and check whether the credits at this tier are genuinely sufficient and whether you keep running short; once you've confirmed it's stable, switch to annual billing to save a good amount (about 47%, check official site for current pricing) — credits are issued per cycle and can be used across all models, not locked to just one.
Self-Check List Before Choosing a Tier
- Have you tested actual consumption with real tasks, rather than estimating by feel?
- Have you accounted for touch-up and local-inpainting costs separately, instead of only counting "first-time generation"?
- Do you have a clear sense of roughly how many images you generate per month?
- Are you treating free credits purely as a way to get a feel for the tool, without expecting them to serve as a long-term production tool?
- After subscribing monthly for a while, do you have a clear sense of whether the credits are enough, before deciding whether to switch to annual?
- Have you factored in your team size, and could everyone share one account instead of each person subscribing separately?
- Have you checked pricing, discounts, and credit numbers against the current official site page, rather than relying on memory or an old screenshot?
- Is upgrading or downgrading flexible enough that you can start from a tier that's just enough, rather than buying the top tier right away?
Being Honest: Free Credits and Lower Tiers Can't Support These Situations
Free credits can't sustain "long-term stable production" — 500 credits (check official site for current terms) is essentially a trial window. Generating a few dozen images to check whether the style suits you is fine, but if you actually try to use it as a daily production tool long-term, you'll sooner or later hit the day your credits run out. That's not the platform being "stingy" — free credits were simply never meant for long-term production use.
The free tier also has a gap in feature ceilings — full speed with no throttling and a higher resolution ceiling are only truly unlocked starting from Pro. The free tier is mainly for verifying results, and expecting it to deliver the exact same stable experience as a paid tier simply isn't realistic.
Exactly how credits are consumed and precisely how many images each tier covers are numbers that can change on the official site — any claim that "one image costs a fixed number of credits" is only a reference point at a specific moment in time. For exact figures, you still need to check the current official site page; guessed numbers don't count.
What annual billing saves you is specifically for long-term stable usage needs. If your own output volume is highly unstable — swamped with orders this month and nothing next month — locking into an annual plan may actually not be the optimal choice. It's better to subscribe monthly first and only consider a long-term tier once your volume stabilizes.