Your rate depends solely on your GPU’s AI performance.

• Base rate: An RTX 4090 earns about $0.32 per full hour at 100 % load.
• Second-by-second billing: Tasks can last from 4 seconds up to 20 minutes—the app pays for the exact runtime.
• Dynamic pricing: If demand rises or a heavier model drops, the rate updates automatically—you always see the live rate before starting a task.

Think of the numbers below like a weather forecast. AI demand changes daily, so real-world earnings can drift up or down.

GPU modelVRAMRelative performance* (vs RTX 4090)Estimated earning* (USD / h)RTX 509032 GB130 %$0.42RTX 508016 GB110 %$0.35RTX 5070 Ti / SUPER18 GB95 %$0.30RTX 409024 GB100 % (baseline)$0.32RTX 408016 GB87 %$0.28

*All figures above are illustrative only. Actual payouts depend on many moving parts—real-time demand for compute, your card’s up-time and VRAM usage, future model updates, and upcoming fee changes. Treat these numbers as a guideline, not a guarantee of income.

At the reference rate of $0.32 per hour for an RTX 4090:

• 1 minute ≈ $0.00533 
• 5 minutes ≈ $0.0267

5 minutes on an RTX 4090 earns roughly $0.027.