diff --git a/docs/cloud-gpus/index.md b/docs/cloud-gpus/index.md index bd499715..f75b9ad6 100644 --- a/docs/cloud-gpus/index.md +++ b/docs/cloud-gpus/index.md @@ -56,7 +56,7 @@ not just inference in particular models as a service. - Direct price comparisons are trickier for serverless offerings: cold boot time and autoscaling logic can substantially impact cost-of-traffic. - Some of the providers allow configuration of CPU and RAM resources. We have selected reasonable defaults, generally comparable to the fixed offerings of other providers. -- You can find pricing pages for the providers here: [Banana](https://banana.dev#pricing), [Baseten](https://docs.baseten.co/settings/pricing), [Beam](https://beam.cloud/pricing), [Modal](https://modal.com/pricing), [Replicate](https://replicate.com/pricing), [RunPod](https://www.runpod.io/serverless-gpu) +- You can find pricing pages for the providers here: [Banana](https://banana.dev#pricing), [Baseten](https://docs.baseten.co/settings/pricing), [Beam](https://beam.cloud/pricing), [Covalent](https://www.covalent.xyz/), [Modal](https://modal.com/pricing), [Replicate](https://replicate.com/pricing), [RunPod](https://www.runpod.io/serverless-gpu) - Serverless GPUs are a newer technology, so the details change quickly and you can expect bugs/growing pains. Stay frosty! - Raw data can be found in a [csv on GitHub](https://github.com/full-stack-deep-learning/website/blob/main/docs/cloud-gpus/serverless-gpus.csv). diff --git a/docs/cloud-gpus/serverless-gpus.csv b/docs/cloud-gpus/serverless-gpus.csv index da592fdc..f50503e9 100644 --- a/docs/cloud-gpus/serverless-gpus.csv +++ b/docs/cloud-gpus/serverless-gpus.csv @@ -12,15 +12,39 @@ Baseten,A100 (80 GB),Ampere,4,320,48,576,24.58,6.15 Baseten,A100 (80 GB),Ampere,8,640,96,1152,49.15,6.14 Beam,A10G (24 GB),Ampere,1,24,4,16,2.11,2.11 Beam,T4 (16 GB),Turing,1,16,4,16,1.55,1.55 -Covalent,T4 (16 GB),Turing,1,16,64,256,0.64,0.64 +Covalent,T4 (16 GB),Turing,1,16,4,16,0.64,0.64 +Covalent,T4 (16 GB),Turing,4,64,48,192,2.56,0.64 Covalent,V100 (16 GB),Volta,1,16,8,61,3.67,3.67 +Covalent,V100 (16 GB),Volta,4,64,32,244,14.68,3.67 +Covalent,V100 (16 GB),Volta,8,128,64,488,29.36,3.67 Covalent,A4000 (16 GB),Ampere,1,16,6,24,0.70,0.70 +Covalent,A4000 (16 GB),Ampere,2,32,12,48,1.40,0.70 +Covalent,A4000 (16 GB),Ampere,4,64,24,96,2.80,0.70 +Covalent,A4000 (16 GB),Ampere,8,128,48,192,5.60,0.70 +Covalent,A4000 (16 GB),Ampere,10,160,64,240,7.00,0.70 Covalent,A5000 (24 GB),Ampere,1,24,8,30,0.90,0.90 +Covalent,A5000 (24 GB),Ampere,2,48,16,60,1.80,0.90 +Covalent,A5000 (24 GB),Ampere,4,96,32,120,3.60,0.90 +Covalent,A5000 (24 GB),Ampere,8,192,64,240,7.20,0.90 Covalent,A6000 (48 GB),Ampere,1,48,16,59.5,1.46,1.46 -Covalent,A10G (24 GB),Ampere,1,24,64,256,1.21,1.21 +Covalent,A6000 (48 GB),Ampere,2,96,32,119,2.92,1.46 +Covalent,A6000 (48 GB),Ampere,4,192,64,238,5.84,1.46 +Covalent,A6000 (48 GB),Ampere,8,384,128,480,11.68,1.46 +Covalent,A10G (24 GB),Ampere,1,24,4,16,1.21,1.21 +Covalent,A10G (24 GB),Ampere,4,96,48,192,4.84,1.21 +Covalent,A10G (24 GB),Ampere,8,192,192,768,9.68,1.21 Covalent,A100 (80 GB),Ampere,1,80,32,120,4.59,4.59 -Covalent,L40 (48 GB),Lovelace,1,48,32,59.5,2.22,2.22 +Covalent,A100 (80 GB),Ampere,2,160,64,240,9.18,4.59 +Covalent,A100 (80 GB),Ampere,4,320,128,480,18.36,4.59 +Covalent,A100 (80 GB),Ampere,8,640,252,960,36.72,4.59 +Covalent,L40 (48 GB),Lovelace,1,48,32,60,2.22,2.22 +Covalent,L40 (48 GB),Lovelace,2,96,64,120,4.44,2.22 +Covalent,L40 (48 GB),Lovelace,4,192,126,240,8.88,2.22 +Covalent,L40 (48 GB),Lovelace,8,384,252,480,17.76,2.22 Covalent,H100 (80 GB),Hopper,1,80,32,180,5.85,5.85 +Covalent,H100 (80 GB),Hopper,2,160,64,360,11.70,5.85 +Covalent,H100 (80 GB),Hopper,4,320,126,720,23.40,5.85 +Covalent,H100 (80 GB),Hopper,8,640,252,1440,46.80,5.85 Modal,T4 (16 GB),Turing,1,16,2,16,1.17,1.17 Modal,A10G (24 GB),Ampere,1,24,4,16,1.87,1.87 Modal,A100 (20 GB),Ampere,1,20,4,16,3.07,3.07