Choose a different AI model for every piece
No single image model is best at everything. artCart.AI connects to the major generators so you can match the model to the artwork instead of fighting the one you happen to be locked into.
You do not need an account with any of these providers, and there is no API key to configure. Pick a model at generation time and pay in credits.
Which model should I use?
Painterly, atmospheric work
Midjourney has a distinctive aesthetic that flatters landscapes, portraits and anything romantic. It returns four variations per run, which is useful when you are exploring rather than executing.
Artwork containing words
Ideogram and GPT Image render legible typography. If the piece is a poster, a quote print or anything with a sign in it, start here.
Long, specific prompts
GPT Image follows detailed instructions closely. Use it when you know exactly what you want and need it honoured.
Photorealism
Gemini 3.1 Flash Image handles natural light, skin and landscape convincingly. Gemini 3 Pro Image pushes further when the result has to survive close inspection.
Large prints
FLUX 1.1 Pro holds fine detail better than most, which matters once you are printing at 24×36 rather than looking at a thumbnail.
Volume and experimentation
Stable Diffusion and FLUX Schnell cost the least per image, so they are the sensible choice when you are generating dozens of options.
Every provider we support
Availability is set by your platform administrator. Models marked Available now can be selected today.
OpenAI
Available nowThe GPT Image family. Excellent prompt adherence and the most legible text in images.
| Model | Credits | Shapes |
|---|---|---|
| GPT Image 2.5 Flare | 10 | 3 ratios |
| GPT Image 2.5 Sunburst | 10 | 3 ratios |
| GPT Image 2 | 5 | 3 ratios |
| GPT Image 1.5 | 4 | 3 ratios |
| GPT Image 1 | 4 | 3 ratios |
| GPT Image 1 mini | 2 | 3 ratios |
Direct: results usually arrive in well under a minute.
Google AI (Gemini)
Available nowGemini image models. Strong photorealism, wide aspect-ratio support and fast turnaround.
| Model | Credits | Shapes |
|---|---|---|
| Gemini 3 Pro Image | 5 | 7 ratios |
| Gemini 3.1 Flash Image | 3 | 7 ratios |
| Gemini 3.1 Flash Lite Image | 2 | 7 ratios |
| Gemini 2.5 Flash Image | 2 | 7 ratios |
Direct: results usually arrive in well under a minute.
Stability AI
Available nowStable Image Ultra, Core and SD 3.5. Broad artistic range at a low per-image cost.
| Model | Credits | Shapes |
|---|---|---|
| Stable Image Ultra | 4 | 7 ratios |
| Stable Image Core | 2 | 7 ratios |
| Stable Diffusion 3.5 | 3 | 7 ratios |
Direct: results usually arrive in well under a minute.
Black Forest Labs (FLUX)
Available nowFLUX 1.1 Pro and FLUX Dev. Very high detail; generation is queued and polled.
| Model | Credits | Shapes |
|---|---|---|
| FLUX 1.1 Pro | 4 | 7 ratios |
| FLUX Pro | 3 | 7 ratios |
| FLUX Dev | 2 | 7 ratios |
Queued: results arrive when the model finishes, typically under a few minutes.
PiAPI (Midjourney)
Midjourney through PiAPI. Returns a grid of four images per run.
| Model | Credits | Shapes |
|---|---|---|
| Midjourney Turbo | 3 ×4 | 7 ratios |
| Midjourney Fast | 2 ×4 | 7 ratios |
| Midjourney Relax | 1 ×4 | 7 ratios |
Queued: results arrive when the model finishes, typically under a few minutes.
Replicate
One key, hundreds of open models. Handy for SDXL, FLUX Schnell and Recraft.
| Model | Credits | Shapes |
|---|---|---|
| FLUX Schnell | 1 | 7 ratios |
| FLUX Dev | 2 | 7 ratios |
| Recraft V3 | 3 | 5 ratios |
| SD 3.5 Large | 3 | 5 ratios |
| Upscale 2x | 1 | 1 ratios |
| Upscale 4x | 2 | 1 ratios |
Queued: results arrive when the model finishes, typically under a few minutes.
xAI (Grok)
Grok image generation. OpenAI-compatible endpoint, up to ten images per run.
| Model | Credits | Shapes |
|---|---|---|
| Grok 2 Image | 2 | 1 ratios |
Direct: results usually arrive in well under a minute.
Ideogram
Best-in-class typography. Ideal for posters and any art with words in it.
| Model | Credits | Shapes |
|---|---|---|
| Ideogram V3 Quality | 4 | 7 ratios |
| Ideogram V3 Balanced | 2 | 7 ratios |
| Ideogram V3 Turbo | 1 | 7 ratios |
Direct: results usually arrive in well under a minute.
fal.ai
Fast hosted runtime for FLUX and friends. Queued API with polling.
| Model | Credits | Shapes |
|---|---|---|
| FLUX Schnell | 1 | 5 ratios |
| FLUX Dev | 2 | 5 ratios |
| FLUX 1.1 Pro | 4 | 5 ratios |
Queued: results arrive when the model finishes, typically under a few minutes.
Credit costs are the platform defaults and may be adjusted by your administrator. Where a model shows a multiplier, it always returns that many images per run.
Questions about models
Do I need my own API key for these?
No. artCart.AI holds the provider connections. You choose a model and spend credits — there is no account to open with OpenAI, Google or anyone else.
Can I switch models halfway through a project?
Yes, every run is independent. Many artists explore with a cheap model and then re-run their best prompt on a more expensive one for the final piece.
Why do some models cost more credits?
The credit price reflects what the provider charges us. Faster queues and higher-quality models cost more per image at the source.
Can I sell what these models produce?
Generally yes, but each provider sets its own commercial-use terms and they do change. Check the terms for the specific model before building a product line on it.
Which model is best for printing large?
FLUX 1.1 Pro and Gemini 3 Pro Image hold the most fine detail. See the model selection guide for a fuller comparison.