Replicate pricing
Replicate bills in two ways. Most public models bill by the time they run on a given GPU. Some models bill by input and output, such as per image or per token. You pay for what you use. Figures come from the Replicate pricing page on October 27, 2026.
Hardware pricing
| Hardware | Per second | Per hour | GPU RAM |
|---|---|---|---|
| CPU (Small) | $0.000025 | $0.09 | None |
| CPU | $0.000100 | $0.36 | None |
| Nvidia T4 | $0.000225 | $0.81 | 16GB |
| Nvidia L40S | $0.000975 | $3.51 | 48GB |
| Nvidia A100 (80GB) | $0.001400 | $5.04 | 80GB |
| Nvidia H100 | $0.001525 | $5.49 | 80GB |
| Nvidia H200 | $0.001525 | $5.49 | Not listed |
| 8x Nvidia H100 | $0.012200 | $43.92 | Not listed |
The page says multi-GPU A100, H100, and L40S capacity and H200 capacity come with committed spend contracts.
Model prices by output
| Model | Price |
|---|---|
| black-forest-labs/flux-1.1-pro | $0.04 per output image |
| black-forest-labs/flux-dev | $0.025 per output image |
| black-forest-labs/flux-schnell | $3.00 per thousand output images |
| ideogram-ai/ideogram-v3-quality | $0.09 per output image |
| recraft-ai/recraft-v3 | $0.04 per output image |
| deepseek-ai/deepseek-r1 | $3.75 per million input tokens, $0.01 per thousand output tokens |
| wavespeedai/wan-2.1-i2v-480p | $0.09 per second of output video |
| wavespeedai/wan-2.1-i2v-720p | $0.25 per second of output video |
Each model page shows a cost estimate.
Private models
Private models you deploy with Cog run on dedicated hardware. You pay for setup time, idle time, and active time. Fast booting fine-tunes are the exception. They bill for active time and not for idle time.
Two quick totals
Our math: a 10-minute run on an A100 (80GB) is 600 seconds at $0.001400, or $0.84. Five FLUX 1.1 Pro images at $0.04 cost $0.20.
What we withhold
We do not list volume discounts. Replicate offers them to large spenders through its enterprise team.
Compare and track
See the fal pricing guide, the Together AI pricing guide, the Runway pricing guide, the Fly.io pricing guide, and all AI tools.
FAQ
How much does Replicate cost? It depends on the model. Time-billed models cost the hardware rate per second. An H100 costs $0.001525 per second, or $5.49 an hour.
How much is an A100 on Replicate? An A100 (80GB) costs $0.001400 per second, or $5.04 an hour.
What is the cheapest GPU on Replicate? The Nvidia T4 at $0.000225 per second, or $0.81 an hour.
How much does FLUX cost on Replicate? flux-1.1-pro costs $0.04 per output image. flux-dev costs $0.025. flux-schnell costs $3.00 per thousand images.
Do private models on Replicate bill for idle time? Yes. Most private models bill for setup, idle, and active time. Fast booting fine-tunes bill for active time.
How we check
We read and rendered the Replicate pricing page on October 27, 2026. Get an alert when Replicate changes a price.
Checked 2026-10-27. Prices and policies change. Confirm on the source before you buy.
Sources
More guides
- 1Password pricing per user: Business plan cost (2026)
- ActiveCampaign pricing: plans, seats and contacts (2026)
- Ahrefs pricing per month: Lite, Standard, Advanced, and Enterprise (2026)
- Aircall pricing per user: Essentials, Professional, and the 3-license minimum (2026)
- Airtable pricing per user: who counts as a paid seat, and team totals (2026)
- Amazon Polly pricing: $4 to $100 per million characters (2026)