Comparisons

All 171 head-to-heads: GPU against GPU on price, memory and published specifications, and provider against provider on the same cards, fees and billing terms. Rates come from today's price checks.

Card against card

Which card costs less per hour, and what you give up for it. Figures: each card's cheapest on-demand rate today.

• A10 vs A30$1.00 vs $0.35/hr
• A10 vs A40$1.00 vs $0.59/hr
• A10 vs L4$1.00 vs $0.44/hr
• A10 vs L40$1.00 vs $0.79/hr
• A10 vs L40S$1.00 vs $0.96/hr
• A10 vs Tesla T4$1.00 vs $0.35/hr
• A10 vs Tesla V100$1.00 vs $0.17/hr
• A100 vs A40$0.89 vs $0.59/hr
• A100 vs B200$0.89 vs $3.49/hr
• A100 vs B300$0.89 vs $4.99/hr
• A100 vs GH200$0.89 vs $2.29/hr
• A100 vs H200$0.89 vs $3.14/hr
• A100 vs L40S$0.89 vs $0.96/hr
• A100 vs MI300X$0.89 vs $2.39/hr
• A100 vs MI325X$0.89 vs $3.80/hr
• A100 vs MI355X$0.89 vs $5.99/hr
• A30 vs A40$0.35 vs $0.59/hr
• A30 vs L4$0.35 vs $0.44/hr
• A30 vs L40$0.35 vs $0.79/hr
• A30 vs L40S$0.35 vs $0.96/hr
• A30 vs Tesla T4$0.35 vs $0.35/hr
• A30 vs Tesla V100$0.35 vs $0.17/hr
• A40 vs L4$0.59 vs $0.44/hr
• A40 vs L40$0.59 vs $0.79/hr
• A40 vs Tesla T4$0.59 vs $0.35/hr
• A40 vs Tesla V100$0.59 vs $0.17/hr
• B200 vs B300$3.49 vs $4.99/hr
• B200 vs GH200$3.49 vs $2.29/hr
• B200 vs H100$3.49 vs $1.99/hr
• B200 vs MI325X$3.49 vs $3.80/hr
• B200 vs MI355X$3.49 vs $5.99/hr
• B300 vs GH200$4.99 vs $2.29/hr
• B300 vs MI300X$4.99 vs $2.39/hr
• B300 vs MI325X$4.99 vs $3.80/hr
• B300 vs MI355X$4.99 vs $5.99/hr
• GH200 vs H100$2.29 vs $1.99/hr
• GH200 vs MI300X$2.29 vs $2.39/hr
• GH200 vs MI325X$2.29 vs $3.80/hr
• GH200 vs MI355X$2.29 vs $5.99/hr
• H100 vs A100$1.99 vs $0.89/hr
• H100 vs B300$1.99 vs $4.99/hr
• H100 vs H200$1.99 vs $3.14/hr
• H100 vs MI325X$1.99 vs $3.80/hr
• H100 vs MI355X$1.99 vs $5.99/hr
• H100 vs RTX 4090$1.99 vs $0.35/hr
• H200 vs B200$3.14 vs $3.49/hr
• H200 vs B300$3.14 vs $4.99/hr
• H200 vs GH200$3.14 vs $2.29/hr
• H200 vs MI300X$3.14 vs $2.39/hr
• H200 vs MI325X$3.14 vs $3.80/hr
• H200 vs MI355X$3.14 vs $5.99/hr
• L4 vs L40$0.44 vs $0.79/hr
• L4 vs L40S$0.44 vs $0.96/hr
• L4 vs Tesla T4$0.44 vs $0.35/hr
• L4 vs Tesla V100$0.44 vs $0.17/hr
• L40 vs L40S$0.79 vs $0.96/hr
• L40 vs Tesla T4$0.79 vs $0.35/hr
• L40 vs Tesla V100$0.79 vs $0.17/hr
• L40S vs A40$0.96 vs $0.59/hr
• L40S vs RTX 4090$0.96 vs $0.35/hr
• L40S vs Tesla T4$0.96 vs $0.35/hr
• L40S vs Tesla V100$0.96 vs $0.17/hr
• MI300X vs B200$2.39 vs $3.49/hr
• MI300X vs H100$2.39 vs $1.99/hr
• MI300X vs MI325X$2.39 vs $3.80/hr
• MI300X vs MI355X$2.39 vs $5.99/hr
• MI325X vs MI355X$3.80 vs $5.99/hr
• RTX 3090 vs RTX 4070 Ti SuperCompared on published specifications
• RTX 3090 vs RTX 4090$0.20 vs $0.35/hr
• RTX 3090 vs RTX 5070 TiCompared on published specifications
• RTX 3090 vs RTX 5080Compared on published specifications
• RTX 3090 vs RTX 5090$0.20 vs $0.78/hr
• RTX 4090 vs A100$0.35 vs $0.89/hr
• RTX 4090 vs RTX 4070 Ti SuperCompared on published specifications
• RTX 4090 vs RTX 5070 TiCompared on published specifications
• RTX 4090 vs RTX 5080Compared on published specifications
• RTX 4090 vs RTX 5090$0.35 vs $0.78/hr
• RTX 5070 Ti vs RTX 4070 Ti SuperCompared on published specifications
• RTX 5080 vs RTX 4070 Ti SuperCompared on published specifications
• RTX 5080 vs RTX 5070 TiCompared on published specifications
• RTX 5090 vs RTX 4070 Ti SuperCompared on published specifications
• RTX 5090 vs RTX 5070 TiCompared on published specifications
• RTX 5090 vs RTX 5080Compared on published specifications
• RTX 6000 Ada vs L40S$0.75 vs $0.96/hr
• RTX 6000 Ada vs RTX PRO 6000$0.75 vs $1.10/hr
• RTX A4000 vs L40S$0.10 vs $0.96/hr
• RTX A4000 vs RTX 6000 Ada$0.10 vs $0.75/hr
• RTX A4000 vs RTX A5000$0.10 vs $0.27/hr
• RTX A4000 vs RTX A6000$0.10 vs $0.35/hr
• RTX A4000 vs RTX PRO 6000$0.10 vs $1.10/hr
• RTX A5000 vs L40S$0.27 vs $0.96/hr
• RTX A5000 vs RTX 6000 Ada$0.27 vs $0.75/hr
• RTX A5000 vs RTX A6000$0.27 vs $0.35/hr
• RTX A5000 vs RTX PRO 6000$0.27 vs $1.10/hr
• RTX A6000 vs L40S$0.35 vs $0.96/hr
• RTX A6000 vs RTX 6000 Ada$0.35 vs $0.75/hr
• RTX A6000 vs RTX PRO 6000$0.35 vs $1.10/hr
• RTX PRO 6000 vs L40S$1.10 vs $0.96/hr
• Rubin vs B200Compared on published specifications
• Rubin vs B300Compared on published specifications
• Rubin vs H200Compared on published specifications
• Tesla V100 vs A100$0.17 vs $0.89/hr
• Tesla V100 vs Tesla T4$0.17 vs $0.35/hr

Provider against provider

Same card, two hosts: rate, fees, and what the billing terms do to the total. Figure: how many GPU models both rent on demand today.

• AWS vs CoreWeave4GPU models both rent
• AWS vs Crusoe4GPU models both rent
• AWS vs DataCrunch4GPU models both rent
• AWS vs Google Cloud4GPU models both rent
• AWS vs Hyperstack3GPU models both rent
• AWS vs Massed Compute4GPU models both rent
• AWS vs Nebius3GPU models both rent
• AWS vs OVHcloud5GPU models both rent
• AWS vs RunPod5GPU models both rent
• CoreWeave vs Crusoe4GPU models both rent
• CoreWeave vs DataCrunch6GPU models both rent
• CoreWeave vs Google Cloud5GPU models both rent
• CoreWeave vs Hyperstack6GPU models both rent
• CoreWeave vs Lambda4GPU models both rent
• CoreWeave vs Massed Compute6GPU models both rent
• CoreWeave vs Nebius5GPU models both rent
• CoreWeave vs OVHcloud4GPU models both rent
• CoreWeave vs RunPod7GPU models both rent
• CoreWeave vs TensorDock3GPU models both rent
• CoreWeave vs Thunder Compute3GPU models both rent
• Crusoe vs DataCrunch4GPU models both rent
• Crusoe vs Google Cloud3GPU models both rent
• Crusoe vs Hyperstack3GPU models both rent
• Crusoe vs Massed Compute4GPU models both rent
• Crusoe vs Nebius3GPU models both rent
• Crusoe vs OVHcloud4GPU models both rent
• Crusoe vs RunPod5GPU models both rent
• DataCrunch vs Google Cloud6GPU models both rent
• DataCrunch vs Hyperstack7GPU models both rent
• DataCrunch vs Lambda5GPU models both rent
• DataCrunch vs Massed Compute7GPU models both rent
• DataCrunch vs Nebius6GPU models both rent
• DataCrunch vs OVHcloud5GPU models both rent
• DataCrunch vs RunPod9GPU models both rent
• DataCrunch vs TensorDock5GPU models both rent
• DataCrunch vs Thunder Compute3GPU models both rent
• Google Cloud vs Hyperstack5GPU models both rent
• Google Cloud vs Lambda4GPU models both rent
• Google Cloud vs Massed Compute4GPU models both rent
• Google Cloud vs Nebius4GPU models both rent
• Google Cloud vs OVHcloud5GPU models both rent
• Google Cloud vs RunPod6GPU models both rent
• Google Cloud vs TensorDock3GPU models both rent
• Hyperstack vs Lambda4GPU models both rent
• Hyperstack vs Massed Compute6GPU models both rent
• Hyperstack vs Nebius5GPU models both rent
• Hyperstack vs OVHcloud3GPU models both rent
• Hyperstack vs RunPod9GPU models both rent
• Hyperstack vs TensorDock5GPU models both rent
• Hyperstack vs Thunder Compute4GPU models both rent
• Lambda vs Massed Compute3GPU models both rent
• Lambda vs OVHcloud4GPU models both rent
• Lambda vs TensorDock4GPU models both rent
• Lambda vs Thunder Compute3GPU models both rent
• Massed Compute vs Nebius4GPU models both rent
• Massed Compute vs OVHcloud4GPU models both rent
• Massed Compute vs RunPod10GPU models both rent
• Massed Compute vs TensorDock5GPU models both rent
• Massed Compute vs Thunder Compute4GPU models both rent
• Nebius vs OVHcloud3GPU models both rent
• Nebius vs RunPod6GPU models both rent
• OVHcloud vs RunPod5GPU models both rent
• OVHcloud vs TensorDock3GPU models both rent
• RunPod vs Lambda4GPU models both rent
• RunPod vs TensorDock8GPU models both rent
• RunPod vs Thunder Compute4GPU models both rent
• RunPod vs Vultr3GPU models both rent
• TensorDock vs Thunder Compute4GPU models both rent