Every GPU we track
54 cards with an on-demand rate card today, most expensive first: the median rate card - each provider once, at its cheapest on-demand rate, community capacity and marketplace floors out - with its panel size, and the cheapest listed on-demand rate with its source and read time. A name opens the card's page, with every listed rate. 24 more are listed only on community capacity or as a marketplace floor: NVIDIA GeForce GTX 1070 Ti, NVIDIA GeForce GTX 1080, NVIDIA GeForce RTX 2060 12GB, NVIDIA GeForce RTX 3060, NVIDIA GeForce RTX 3060 Ti, NVIDIA GeForce RTX 3070, NVIDIA GeForce RTX 3070 Ti, NVIDIA GeForce RTX 3080 Ti, NVIDIA GeForce RTX 4060, NVIDIA GeForce RTX 4060 Ti, NVIDIA GeForce RTX 4070, NVIDIA GeForce RTX 4070 SUPER, NVIDIA GeForce RTX 4070 Ti, NVIDIA GeForce RTX 4080 SUPER, NVIDIA GeForce RTX 5060, NVIDIA GeForce RTX 5060 Ti, NVIDIA GeForce RTX 5070, NVIDIA GeForce RTX 5070 Ti, NVIDIA Quadro P4000, NVIDIA RTX A2000, NVIDIA Tesla P100 12GB, NVIDIA TITAN RTX, NVIDIA TITAN V, NVIDIA TITAN Xp.
Today’s GPU-hour
Median rate card per card: each provider once, at its cheapest on-demand rate, community capacity and marketplace floors out — the card median every page on this site prints. The H100 SXM is headed by the Price Index instead, which has rules of its own (USD listings only). $/GB divides it by the card’s memory; spread is the dearest rate over the cheapest, with a dot per provider between them. Tap a card to open its page.
Memory| Card, $ per GB, spread and providers | Where each provider sits between cheapest and dearest | Median rate card, cheapest and dearest |
|---|---|---|
| GB300 NVL72288GB$0.063/GB · 1.6× spread · 3 providers | $18.00$12.54–$20.007d +2.0% | |
| GB200 NVL72186GB$0.071/GB · 3.9× spread · 6 providers | $13.25$6.94–$27.047d no change | |
| B300288GB$0.031/GB · 3.0× spread · 25 providers | $8.99$5.89–$17.807d +0.1% | |
| B200192GB$0.036/GB · 3.9× spread · 39 providers | $6.99$4.11–$16.117d +0.1% | |
| MI355X288GB$0.021/GB · 2.9× spread · 3 providers | $5.99$2.95–$8.607d no change | |
| MI350X288GB$0.018/GB · 1.5× spread · 2 providers | $5.08$4.00–$6.167d no change | |
| H200141GB$0.032/GB · 5.1× spread · 56 providers | $4.54$2.09–$10.607d −1.5% | |
| H100 NVL94GB$0.043/GB · 2.8× spread · 8 providers | $4.07$2.92–$8.037d +1.9% | |
| H100 80GB (form factor not stated)80GB$0.049/GB · 5.1× spread · 27 providers | $3.95$2.10–$10.617d +1.2% | |
| H200 NVL141GB$0.026/GB · 1.7× spread · 6 providers | $3.71$3.36–$5.607d no change | |
| H100 SXM80GB$0.044/GB · 6.2× spread · 41 providers | $3.54$1.79–$11.067d no change | |
| GH20096GB$0.033/GB · 2.8× spread · 4 providers | $3.20$2.29–$6.507d no change | |
| H100 PCIe80GB$0.035/GB · 3.7× spread · 12 providers | $2.82$1.83–$6.827d no change | |
| MI300X192GB$0.015/GB · 3.5× spread · 8 providers | $2.81$1.71–$6.007d no change | |
| MI325X256GB$0.009/GB · 1.7× spread · 3 providers | $2.25$2.25–$3.807d no change | |
| RTX PRO 6000 Blackwell96GB$0.023/GB · 8.5× spread · 41 providers | $2.23$0.65–$5.507d −0.6% | |
| A100 80GB80GB$0.025/GB · 6.5× spread · 63 providers | $1.99$1.05–$6.827d +1.5% | |
| A100 40GB40GB$0.050/GB · 5.4× spread · 22 providers | $1.98$0.63–$3.407d +0.2% | |
| RTX PRO 6000 Blackwell Max-Q96GB$0.019/GB · 1.1× spread · 2 providers | $1.86$1.80–$1.927d no change | |
| V10016GB$0.103/GB · 13.2× spread · 12 providers | $1.65$0.23–$3.067d +0.2% | |
| L40S48GB$0.032/GB · 5.6× spread · 42 providers | $1.52$0.63–$3.547d +3.0% | |
| A1024GB$0.051/GB · 9.4× spread · 9 providers | $1.22$0.34–$3.207d no change | |
| RTX PRO 5000 Blackwell48GB$0.025/GB · 1.5× spread · 2 providers | $1.21$0.96–$1.457d no change | |
| Intel Gaudi 296GB$0.012/GB · 1.1× spread · 2 providers | $1.16$1.10–$1.21 | |
| V100 32GB32GB$0.036/GB · 20.3× spread · 11 providers | $1.15$0.16–$3.257d no change | |
| RTX 6000 Ada48GB$0.022/GB · 2.8× spread · 11 providers | $1.06$0.65–$1.827d +0.5% | |
| A10G24GB$0.044/GB · 1.7× spread · 5 providers | $1.05$1.00–$1.677d no change | |
| L4048GB$0.020/GB · 1.9× spread · 11 providers | $0.97$0.78–$1.507d no change | |
| RTX PRO 4500 Blackwell32GB$0.027/GB · 3.5× spread · 6 providers | $0.86$0.72–$2.527d no change | |
| RTX 5000 Ada Generation32GB1.0× spread · 1 provider | one rate card, no median$0.83–$0.83 | |
| RTX 509032GB$0.025/GB · 2.9× spread · 20 providers | $0.81$0.55–$1.587d −0.3% | |
| L424GB$0.033/GB · 4.3× spread · 24 providers | $0.80$0.27–$1.177d no change | |
| GB10 (DGX Spark)128GB$0.006/GB · 1.1× spread · 2 providers | $0.77$0.75–$0.79 | |
| Quadro RTX 40008GB1.0× spread · 1 provider | one rate card, no median$0.65–$0.65 | |
| A4048GB$0.013/GB · 5.2× spread · 8 providers | $0.64$0.39–$2.047d no change | |
| RTX A600048GB$0.012/GB · 7.1× spread · 21 providers | $0.59$0.35–$2.507d +0.6% | |
| RTX A500024GB$0.024/GB · 5.8× spread · 10 providers | $0.58$0.27–$1.567d no change | |
| T416GB$0.036/GB · 4.1× spread · 18 providers | $0.57$0.24–$0.997d +1.9% | |
| RTX 4000 Ada20GB$0.026/GB · 3.1× spread · 5 providers | $0.52$0.28–$0.877d no change | |
| Quadro RTX 600024GB$0.021/GB · 5.6× spread · 4 providers | $0.51$0.27–$1.507d no change | |
| GeForce RTX 508016GB$0.030/GB · 1.6× spread · 2 providers | $0.48$0.36–$0.597d no change | |
| RTX 4000 SFF Ada20GB$0.024/GB · 1.1× spread · 2 providers | $0.47$0.44–$0.507d no change | |
| RTX 409024GB$0.019/GB · 3.0× spread · 16 providers | $0.47$0.33–$0.997d −5.0% | |
| GeForce RTX 3090 Ti24GB1.0× spread · 1 provider | one rate card, no median$0.46–$0.46 | |
| A3024GB$0.017/GB · 6.1× spread · 9 providers | $0.41$0.22–$1.357d no change | |
| GeForce RTX 309024GB$0.017/GB · 3.7× spread · 9 providers | $0.40$0.21–$0.787d −4.9% | |
| GeForce RTX 308010GB$0.035/GB · 2.0× spread · 2 providers | $0.35$0.23–$0.477d no change | |
| GeForce RTX 408016GB1.0× spread · 1 provider | one rate card, no median$0.30–$0.30 | |
| GeForce RTX 4070 Ti SUPER16GB1.0× spread · 1 provider | one rate card, no median$0.29–$0.29 | |
| GeForce RTX 2080 Ti11GB$0.023/GB · 1.2× spread · 2 providers | $0.26$0.23–$0.297d no change | |
| RTX 2000 Ada16GB$0.015/GB · 1.0× spread · 2 providers | $0.24$0.24–$0.247d no change | |
| RTX A400016GB$0.015/GB · 9.2× spread · 11 providers | $0.23$0.12–$1.107d +22.5% | |
| A216GB$0.015/GB · 1.4× spread · 2 providers | $0.23$0.20–$0.277d no change | |
| GeForce GTX 1080 Ti11GB1.0× spread · 1 provider | one rate card, no median$0.23–$0.23 |
By workload - which cards hold an 8B or a 70B model, a fine-tune, an image model - and the cheapest that fits: which GPU for which workload.
Cheapest listed rate, with its source
| Card, memory | Median rate card | Cheapest on-demand |
|---|---|---|
| NVIDIA GB300 NVL72 · 288 GB | $18.00 3 providers | $12.54  Verdacaptured |
| NVIDIA GB200 NVL72 · 186 GB | $13.25 6 providers | $6.94  Bitdeer AI · fromcaptured |
| NVIDIA B300 · 288 GB | $8.99 25 providers | $5.89  Deep Infracaptured |
| NVIDIA B200 · 192 GB | $6.99 39 providers | $4.11  Beam · fromcaptured |
| AMD Instinct MI355X · 288 GB | $5.99 3 providers | $2.95  TensorWave · fromcaptured |
| AMD Instinct MI350X · 288 GB | $5.08 2 providers | $4.00  CloudRiftcaptured |
| NVIDIA H200 · 141 GB | $4.54 56 providers | $2.09 Beam · from captured |
| NVIDIA H100 NVL · 94 GB | $4.07 8 providers | $2.92  Massed Computecaptured |
| NVIDIA H100 80GB (form factor not stated) · 80 GB | $3.95 27 providers | $2.10 FlexAI · GPU only captured |
| NVIDIA H200 NVL · 141 GB | $3.71 6 providers | $3.36  IONOS Cloudcaptured |
| NVIDIA H100 SXM · 80 GB | $3.54 41 providers | $1.79  Cudo Compute · GPU onlycaptured |
| NVIDIA GH200 · 96 GB | $3.20 4 providers | $2.29  Lambdacaptured |
| NVIDIA H100 PCIe · 80 GB | $2.82 12 providers | $1.83 Beam · from captured |
| AMD Instinct MI300X · 192 GB | $2.81 8 providers | $1.71 TensorWave · from captured |
| AMD Instinct MI325X · 256 GB | $2.25 3 providers | $2.25  Bentauscaptured |
| NVIDIA RTX PRO 6000 Blackwell · 96 GB | $2.23 41 providers | $0.65  Hostinger · fromcaptured |
| NVIDIA A100 80GB · 80 GB | $1.99 63 providers | $1.05 CloudRift captured |
| NVIDIA A100 40GB · 40 GB | $1.98 22 providers | $0.63  CloudGPUcaptured |
| NVIDIA RTX PRO 6000 Blackwell Max-Q · 96 GB | $1.86 2 providers | $1.80  EmpirioLabscaptured |
| NVIDIA V100 · 16 GB | $1.65 12 providers | $0.23 Verda captured |
| NVIDIA L40S · 48 GB | $1.52 42 providers | $0.63 CloudRift captured |
| NVIDIA A10 · 24 GB | $1.22 9 providers | $0.34  Intelion Cloud · GPU onlycaptured |
| NVIDIA RTX PRO 5000 Blackwell · 48 GB | $1.21 2 providers | $0.96  RunPodcaptured |
| Intel Gaudi 2 · 96 GB | $1.16 2 providers | $1.10  Denvr Dataworkscaptured |
| NVIDIA V100 32GB · 32 GB | $1.15 11 providers | $0.16  Geoddcaptured |
| NVIDIA RTX 6000 Ada · 48 GB | $1.06 11 providers | $0.65  GPU.ai · fromcaptured |
| NVIDIA A10G · 24 GB | $1.05 5 providers | $1.00  Hugging Facecaptured |
| NVIDIA L40 · 48 GB | $0.97 11 providers | $0.78 GPU.ai · from captured |
| NVIDIA RTX PRO 4500 Blackwell · 32 GB | $0.86 6 providers | $0.72 GPU.ai · from captured |
| NVIDIA RTX 5000 Ada Generation · 32 GB | — one rate card, no median | $0.83 RunPod captured |
| NVIDIA RTX 5090 · 32 GB | $0.81 20 providers | $0.55  HexGridcaptured |
| NVIDIA L4 · 24 GB | $0.80 24 providers | $0.27  Hikube · GPU onlycaptured |
| NVIDIA GB10 (DGX Spark) · 128 GB | $0.77 2 providers | $0.75  Envergecaptured |
| NVIDIA Quadro RTX 4000 · 8 GB | — one rate card, no median | $0.65  airgpucaptured |
| NVIDIA A40 · 48 GB | $0.64 8 providers | $0.39 CloudGPU captured |
| NVIDIA RTX A6000 · 48 GB | $0.59 21 providers | $0.35  Thunder Computecaptured |
| NVIDIA RTX A5000 · 24 GB | $0.58 10 providers | $0.27 GPU.ai · from captured |
| NVIDIA T4 · 16 GB | $0.57 18 providers | $0.24  immers.cloudcaptured |
| NVIDIA RTX 4000 Ada · 20 GB | $0.52 5 providers | $0.28 GPU.ai · from captured |
| NVIDIA Quadro RTX 6000 · 24 GB | $0.51 4 providers | $0.27  centron · GPU onlycaptured |
| NVIDIA GeForce RTX 5080 · 16 GB | $0.48 2 providers | $0.36  SwissGPUcaptured |
| NVIDIA RTX 4000 SFF Ada · 20 GB | $0.47 2 providers | $0.44 RunPod captured |
| NVIDIA RTX 4090 · 24 GB | $0.47 16 providers | $0.33 CloudGPU captured |
| NVIDIA GeForce RTX 3090 Ti · 24 GB | — one rate card, no median | $0.46 RunPod captured |
| NVIDIA A30 · 24 GB | $0.41 9 providers | $0.22 CloudGPU captured |
| NVIDIA GeForce RTX 3090 · 24 GB | $0.40 9 providers | $0.21 CloudGPU captured |
| NVIDIA GeForce RTX 3080 · 10 GB | $0.35 2 providers | $0.23  SynpixCloudcaptured |
| NVIDIA GeForce RTX 4080 · 16 GB | — one rate card, no median | $0.30 SwissGPU captured |
| NVIDIA GeForce RTX 4070 Ti SUPER · 16 GB | — one rate card, no median | $0.29  Trooper.AI · fromcaptured |
| NVIDIA GeForce RTX 2080 Ti · 11 GB | $0.26 2 providers | $0.23 SynpixCloud captured |
| NVIDIA RTX 2000 Ada · 16 GB | $0.24 2 providers | $0.24 GPU.ai · from captured |
| NVIDIA RTX A4000 · 16 GB | $0.23 11 providers | $0.12 GPU.ai · from captured |
| NVIDIA A2 · 16 GB | $0.23 2 providers | $0.20 Infomaniak captured |
| NVIDIA GeForce GTX 1080 Ti · 11 GB | — one rate card, no median | $0.23 SynpixCloud captured |
Cards compared
Pairs of cards weighed against each other: memory, the median and cheapest listed rental rate per GPU-hour, price per GB of VRAM, and which providers list both. This list of 14 is curated rather than generated — every combination of 54 models would be noise.
| Comparison | Median rate card, first | Median rate card, second | VRAM |
|---|---|---|---|
| NVIDIA H100 SXM vs NVIDIA H100 PCIe | $3.54 | $2.82 | 80 vs 80 GB |
| NVIDIA H100 SXM vs NVIDIA H200 | $3.54 | $4.54 | 80 vs 141 GB |
| NVIDIA H200 vs NVIDIA B200 | $4.54 | $6.99 | 141 vs 192 GB |
| NVIDIA B200 vs NVIDIA B300 | $6.99 | $8.99 | 192 vs 288 GB |
| NVIDIA GB200 NVL72 vs NVIDIA B200 | $13.25 | $6.99 | 186 vs 192 GB |
| NVIDIA A100 80GB vs NVIDIA H100 SXM | $1.99 | $3.54 | 80 vs 80 GB |
| NVIDIA A100 40GB vs NVIDIA A100 80GB | $1.98 | $1.99 | 40 vs 80 GB |
| NVIDIA H100 SXM vs NVIDIA B200 | $3.54 | $6.99 | 80 vs 192 GB |
| NVIDIA L40S vs NVIDIA RTX A6000 | $1.52 | $0.59 | 48 vs 48 GB |
| NVIDIA RTX 4090 vs NVIDIA RTX PRO 6000 Blackwell | $0.47 | $2.23 | 24 vs 96 GB |
| NVIDIA RTX PRO 6000 Blackwell vs NVIDIA L40S | $2.23 | $1.52 | 96 vs 48 GB |
| NVIDIA RTX 4090 vs NVIDIA L40S | $0.47 | $1.52 | 24 vs 48 GB |
| NVIDIA H100 SXM vs AMD Instinct MI300X | $3.54 | $2.81 | 80 vs 192 GB |
| NVIDIA H200 vs AMD Instinct MI300X | $4.54 | $2.81 | 141 vs 192 GB |
Comparing providers rather than cards: provider comparisons. Every model’s own page carries its full price table and history.
Data updated: