GPU Cloud Pricing
Compare observed GPU rental prices (Vast.ai) · Updated Oct 30, 2026
19
19 total offers
$0.48/hr
$0.016/hr
RTX 3080
N/A
no perf data
N/A
Cheapest option across Vast.ai, RunPod, TensorDock & Lambda Labs
4x RTX 3080 (40GB total)
via Vast.ai
$0.062/hr
2x RTX 4090D (48GB total)
via Vast.ai
$0.216/hr
2x RTX 4090 (48GB total)
via Vast.ai
$0.243/hr
2x RTX 3090 (48GB total)
via Vast.ai
$0.244/hr
3x RTX 5070 Ti (48GB total)
via Vast.ai
$0.302/hr
Daily minimum across all providers (Vast.ai, RunPod, TensorDock, Lambda Labs). Flat lines are normal: the cheapest listing rarely changes — the minimum only moves when a cheaper machine appears. Series rebuilt clean on 2026-09-03 from all 4 providers, refreshed every night.
Live offers — cheapest first
| GPU | Provider | VRAM | $/hour | Region | Rent |
|---|---|---|---|---|---|
| RTX 3080cheapest | Vast.ai | 10 GB | $0.016 | Ontario, CA | rent |
| RTX 5070 Ti | Vast.ai | 16 GB | $0.101 | Vietnam, VN | rent |
| RTX 5070 | Vast.ai | 12 GB | $0.104 | Texas, US | rent |
| RTX 4090D | Vast.ai | 24 GB | $0.108 | , CN | rent |
| RTX 4070S Ti | Vast.ai | 16 GB | $0.108 | Oklahoma, US | rent |
| RTX 4090 | Vast.ai | 24 GB | $0.121 | , CN | rent |
| RTX 3090 | Vast.ai | 24 GB | $0.122 | Taiwan, TW | rent |
| RTX 5080 | Vast.ai | 16 GB | $0.135 | Yunnan, CN | rent |
| RTX 5060 Ti | Vast.ai | 16 GB | $0.205 | California, US | rent |
| RTX 5090 | Vast.ai | 32 GB | $0.213 | South Korea, KR | rent |
| L40S | Vast.ai | 45 GB | $0.336 | Virginia, US | rent |
| RTX PRO 4500 | Vast.ai | 32 GB | $0.349 | Belgium, BE | rent |
| A800 PCIE | Vast.ai | 80 GB | $0.481 | Hunan, CN | rent |
| RTX 6000Ada | Vast.ai | 48 GB | $0.539 | California, US | rent |
| RTX PRO 5000 | Vast.ai | 48 GB | $0.562 | Georgia, US | rent |
| RTX PRO 6000 WS | Vast.ai | 96 GB | $0.736 | Arizona, US | rent |
| RTX PRO 6000 S | Vast.ai | 96 GB | $1.244 | Quebec, CA | rent |
| RTX PRO 6000 Max-Q | Vast.ai | 96 GB | $1.274 | Zhejiang, CN | rent |
| H100 NVL | Vast.ai | 94 GB | $2.336 | Iowa, US | rent |
19 of 19 live offers · cheapest first · one row per provider offer. Vast.ai is a peer-to-peer marketplace — usually cheapest on consumer GPUs (RTX, V100). TensorDock, RunPod and Lambda are datacenter providers — often cheaper on A100/H100-class.
Data sourced from the Vast.ai API only · Prices in USD · Updated nightly · Other providers removed 20/09/2026 (their prices were hard-coded and dated 2024-12-15)