How is total time estimated?

The planner estimates seconds per image from the model, GPU, resolution, and step count, then multiplies by your image count. It is a planning estimate — real runs vary with sampler, VRAM headroom, and whether you batch images per call.

When is cloud cheaper than local?

Cloud wins for occasional large bursts where you do not own a fast GPU, or when you need many images quickly and can rent a high-end card by the hour. If you generate constantly, owning a 3090/4090 amortises fast and local becomes far cheaper per image.

Does a higher step count always look better?

No. Modern samplers plateau around 20–30 steps for most models; beyond that you mostly burn time and credits. Drop steps first when you need to cut a large batch's cost.

How do I cut the cost of a big batch?

Lower resolution and upscale the keepers afterwards, reduce steps to the sampler's sweet spot, generate a small test batch to fix the prompt before committing, and batch multiple images per GPU call to amortise model load.

What is the Batch Image Generation Planner?

Enter batch parameters — image count, resolution, steps, model, and hardware — to estimate total generation time, cloud credit cost, and whether local or cloud generation is the better choice for your job. It runs free in your browser on Gera Tools, with nothing uploaded.

Batch Image Generation Planner

Name: Batch Image Generation Planner
Creator: Gera Tools
License: https://creativecommons.org/licenses/by/4.0/

Batch image generation planning

Generating one image is cheap; generating two thousand for a product catalog, game assets, or a dataset is a real time-and-money decision. This planner estimates the total time, the cloud credit cost, and whether local or cloud hardware is the smarter call for your batch.

When to use this planner

The gap between a casual generation session and a production batch is significant. A concept artist exploring ideas will generate dozens of images; a training data pipeline, a product photography replacement workflow, or a game asset factory may need tens of thousands. At that scale, decisions about resolution, step count, and whether to rent GPUs can mean the difference between a two-hour job and a two-day one — or between a £20 bill and a £200 one. Use this planner before committing to a large run.

How it works

The core estimate chains three numbers:

sec_per_image  = base(model, gpu) × res_scale × step_scale
total_time     = sec_per_image × image_count
cloud_cost     = (total_time / 3600) × gpu_hourly_rate

Resolution scales with pixel count and step count scales linearly, mirroring how diffusion actually works (each step is one UNet pass). For cloud, the planner applies a representative hourly rate for the GPU tier; for local, the marginal electricity cost is small, so the recommendation hinges on whether you already own the card.

Tips for cheaper, faster batches

Prototype small. Lock your prompt and seed range on 4–8 images before launching thousands.
Right-size resolution. Generate at the model’s native res and upscale only the keepers — cheaper than generating everything large.
Trim steps to the sweet spot. 20–30 steps is plenty for most samplers; extra steps mostly cost money.
Batch per call. Generating 4–8 images per GPU call amortises model load and raises throughput on cloud instances.

Local vs cloud: the real trade-off

For a one-off batch of a few hundred images, cloud GPU rental is often the most practical choice even if you own a capable local card — you avoid tying up your machine for hours and can parallelise across multiple cloud instances. The break-even point shifts decisively toward local hardware if you generate regularly: a modern high-end consumer GPU purchased outright typically pays back its cost in cloud-equivalent savings over six to twelve months of moderate use.

The planner flags when your batch is large enough that cloud costs reach a meaningful threshold, and when a local card would complete the job in a comparable time window. Neither option dominates the other — it depends on your generation cadence, the GPU you own, and whether the batch is time-sensitive.

Understanding the step-count trade-off

Each diffusion step is one forward pass through the UNet, so cost scales linearly with steps. The quality improvement per additional step is not linear, though — most of the visible quality gain happens in the first 15–20 steps for well-tuned samplers (DPM++ 2M, Euler a, and their variants). Steps beyond 30 add runtime cost while producing diminishing returns visible to most viewers. For production batches, testing at 20 steps versus 28 steps on a small proof-of-concept and evaluating acceptability is worth the upfront time: cutting steps from 30 to 20 reduces total batch time by one-third with no additional change to any other setting.