Skip to content
Superluminal

Blender render farm pricing

By Superluminal Engineering

Superluminal bills the measured render time of each finished frame. Twelve benchmark jobs cost $0.14 to $1.90 each.

Published 2026-07-29 · Updated 2026-08-06

Billed work
Finished frame task

A frame task adds a charge only after it finishes.

Task charge
runtime × GPUs on node × current rate

Runtime is converted from seconds to hours.

Not billed
Queueing, failed or canceled attempts, and delivery

Finished frames from before cancellation remain billed.

Twelve measured jobs
$9.56 total

The reconciled charges ranged from $0.14 to $1.90.

Superluminal charges the measured render time of each finished frame. Queue time, startup, downloads, failed tasks, and canceled tasks add no charge.

Frames that finish before a cancellation remain billed.

Choose capacity in GPUs and nodes

The capacity control selects the total GPU count. It also lets you place one, two, or three GPUs on each node.

The control shows the current rate for the selected GPU count. More GPUs can finish independent frames sooner, but the per-GPU rate also rises.

Superluminal billing page set to 8 GPUs at $2.26 per GPU-hour
The capacity control shows the selected GPU count, current rate, and live allocation. Open image

The live rate table supplies current prices, and the dated rate dataset records the curve below.

Published rate anchors by selected GPU count

The dated curve lists the per-GPU-hour rate at four capacity anchors. The live table supplies the exact rate for each integer GPU count.

Superluminal rate per GPU-hour rises from 1 dollar at one GPU to 4 dollars from 30 GPUs
Rates captured on August 6, 2026. Download chart PNG

The billing guide explains the same control in the web app. The calculator lets you test any count from one to 30 nodes.

Calculate one frame charge

Each finished frame adds this charge:

$ \text{frame charge}=\frac{\text{render seconds}}{3600}\times\text{GPUs on the node}\times\text{rate per GPU-hour} $

The rate comes from the total GPU capacity active for your organization. Billing uses the successful frame runtime measured on the farm.

Running Superluminal job summary with accrued, estimated, per-frame, and hourly costs
The job summary separates accrued cost, estimated total, per-frame cost, and hourly rate. Open image

The job page updates the charge as frames finish. See job details for the live cost and compute fields.

Measured prices at three capacities

The benchmark rendered four scenes at five, ten, and 30 selected nodes. These are the exact job charges shown by Superluminal.

Recorded render cost by Blender scene

Each bar totals the recorded USD-qualified charges across the four scene runs for one tested setup. A charged retrieval-defect run remains cost evidence. Superluminal includes its 5, 10, and 30-node setups; iRender is omitted because its recordings expose only native iR estimates, not normalized USD.

Seven Blender render farm setups compared by total recorded USD-qualified cost across four scene runs
Totals across all four scenes. Download chart PNG
Scene5 nodes10 nodes30 nodes
Classroom, 145 Cycles frames$1.10$1.35$1.90
Junk Shop, 96 Cycles frames$0.48$0.64$1.21
Monster Under the Bed, 96 Cycles frames$0.50$0.62$0.98
Fox, 50 EEVEE frames$0.14$0.33$0.31
Four-scene total$2.22$2.94$4.40

Five nodes cost the least for every scene. Thirty nodes had the shortest render pipeline on the three Cycles scenes, and ten nodes were fastest on the Fox scene.

In the render farm benchmark, no tested plan charged less than Superluminal on any scene.

Estimate your animation

Render one representative frame on your machine. Enter its time, your hardware, and the frame count in the Blender render cost calculator.

The estimate uses the selected node count and the same public rate table as the capacity control. Your job page remains the source for the final charge.

Sources

  1. Superluminal public pricing endpoint
  2. Superluminal render farm payments and refunds policy
  3. Superluminal pricing reference
  4. Blender render cost model
  5. Render farm benchmark dataset, July to August 2026