Shello
Solana: readingReading the network... See the network

20 m

The shell market

The cheapest H100, the GPU most big AI models run on, costs $0.890 an hour right now.

Shello with sparkling eyes, claws up

Nosana: io.net: Akash: Vast.ai: off (its terms forbid scripted reads)

25 GPUs with a live priceVRAM is the GPU's own memory. A model must fit in it: about 0.5 GB per billion parameters at 4-bit, plus a few GB to run. More in the glossary

cheapest live offer per card, per hour
B300$4.48/hour288 GB memory on io.netfits about a 520B model at 4-bit
H100$0.890/hour80 GB memory on io.netfits about a 140B model at 4-bitcloud list from $1.99under half the cloud price
A100 80GB$0.900/hour80 GB memory on Nosanafits about a 140B model at 4-bitcloud list from $1.19
A100 40GB$0.610/hour40 GB memory on Nosanafits about a 65B model at 4-bitcloud list from $1.99under half the cloud price
RTX PRO 6000$2.07/hour96 GB memory on Akashfits about a 170B model at 4-bit
L40S$0.790/hour48 GB memory on io.netfits about a 80B model at 4-bitcloud list from $0.790
L40$0.790/hourmemory not stated on io.net
RTX 6000 Ada$0.700/hour48 GB memory on Nosanafits about a 80B model at 4-bitcloud list from $0.740
RTX A6000$0.320/hour48 GB memory on io.netfits about a 80B model at 4-bitcloud list from $0.330
A40$0.400/hourmemory not stated on Nosana
RTX 5090$0.500/hour32 GB memory on Nosanafits about a 50B model at 4-bitcloud list from $0.690
RTX 4090$0.250/hour24 GB memory on io.netfits about a 40B model at 4-bitcloud list from $0.340
RTX 3090$0.192/hour24 GB memory on Nosanafits about a 40B model at 4-bitcloud list from $0.220
RTX 5080$0.200/hourmemory not stated on Nosana
RTX 4080$0.160/hourmemory not stated on Nosana
RTX A5000$0.200/hourmemory not stated on Nosana
RTX 3080$0.096/hourmemory not stated on Nosana
RTX 4070$0.096/hourmemory not stated on Nosana
RTX 3070$0.080/hourmemory not stated on Nosana
RTX 3060$0.048/hourmemory not stated on Nosana
T4$0.050/hour16 GB memory on io.netfits about a 25B model at 4-bit
V100$0.200/hour32 GB memory on Akashfits about a 50B model at 4-bit
RTX 5070$0.150/hourmemory not stated on Nosana
RTX 4060$0.064/hourmemory not stated on Nosana
RTX A4000$0.128/hourmemory not stated on Nosana
Show every offer (114, for developers)
$ / GPU-h
dollars for one GPU for one hour
free / total
GPUs of that card free now, out of all listed
read
UTC time of the live read, or the day a cloud list price was read
cardwhereoffer$ / GPU-hfree / totalread
B300io.netB300 SXM6 AC$4.480 / 806:02:07 UTC
B200Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$5.98-2026-11-09
Cloud listLambda, not stated$7.69-2026-11-09
H200Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$3.59-2026-11-09
Cloud listDigitalOcean (GPU Droplets), not stated$4.47-2026-11-09
H100io.netH100 PCIe$0.8900 / 806:02:07 UTC
io.netH100 80GB HBM3$1.190 / 96006:02:07 UTC
A100 80GBNosanaNVIDIA A100 80GB (premium)$0.9000 / 1306:01:53 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$1.19-2026-11-09
A100 40GBNosanaNVIDIA A100 40GB (premium)$0.6100 / 006:01:53 UTC
Akashnvidia a100 40Gi SXM4 (median of listings, lowest $0.84)$0.8400 / 106:02:01 UTC
RTX PRO 6000Akashnvidia pro6000se 96Gi PCIe (median of listings, lowest $1.52)$2.0733 / 8106:02:01 UTC
NosanaNVIDIA Pro 6000 (SOC2) (premium)$2.240 / 206:01:53 UTC
L40Sio.netL40S$0.7900 / 806:02:07 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.790-2026-11-09
L40io.netL40$0.7900 / 706:02:07 UTC
RTX 6000 AdaNosanaNVIDIA 6000 Ada (premium)$0.7002 / 206:01:53 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.740-2026-11-09
RTX A6000io.netRTX A6000$0.3200 / 106:02:07 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.330-2026-11-09
A40NosanaNVIDIA A40 (premium)$0.4000 / 106:01:53 UTC
RTX 5090NosanaNVIDIA 5090 (premium)$0.5000 / 4606:01:53 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.690-2026-11-09
RTX 4090io.netGeForce RTX 4090$0.2500 / 806:02:07 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.340-2026-11-09
RTX 3090NosanaNVIDIA 3090 (premium)$0.19212 / 3506:01:53 UTC
Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.220-2026-11-09
RTX 5080NosanaNVIDIA 5080 (premium)$0.2005 / 1006:01:53 UTC
RTX 4080NosanaNVIDIA 4080 (premium)$0.1600 / 606:01:53 UTC
RTX A5000NosanaNVIDIA 5000/A5000 (premium)$0.2000 / 106:01:53 UTC
A10Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$1.01-2026-11-09
Cloud listLambda, not stated$1.29-2026-11-09
L4Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$0.440-2026-11-09
Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$0.590-2026-11-09
RTX 3080NosanaNVIDIA 3080 (premium)$0.09614 / 2506:01:53 UTC
RTX 4070NosanaNVIDIA 4070 (premium)$0.0960 / 806:01:53 UTC
RTX 3070NosanaNVIDIA 3070 (premium)$0.0803 / 906:01:53 UTC
RTX 3060NosanaNVIDIA 3060 (premium)$0.0487 / 1406:01:53 UTC
T4io.netTesla T4$0.0500 / 306:02:07 UTC
Akashnvidia t4 16Gi PCIe (median of listings, lowest $0.11)$0.1501 / 206:02:01 UTC
V100Akashnvidia v100 32Gi PCIe (median of listings, lowest $0.2)$0.2001 / 406:02:01 UTC
RTX 5070NosanaNVIDIA 5070 (premium)$0.1507 / 2206:01:53 UTC
RTX 4060NosanaNVIDIA 4060 (premium)$0.0641 / 406:01:53 UTC
RTX A4000NosanaNVIDIA 4000/A4000 (premium)$0.1280 / 106:01:53 UTC
H100 NVLCloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$2.59-2026-11-09
Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$3.19-2026-11-09
See all 114 offers
cardwhereoffer$ / GPU-hread
B200Cloud listLambda, not stated$7.992026-11-09
B200Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$7.992026-11-09
B200Cloud listTogether AI (GPU Clusters, on-demand), not stated$8.192026-11-09
B200Cloud listNebius, not stated$8.502026-11-09
B200Cloud listCoreWeave, North America$8.602026-11-09
B200Cloud listFireworks (on-demand deployments), not stated; region-restricted deployments 1.5x$13.002026-11-09
B200Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$14.242026-11-09
H200Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$5.292026-11-09
H200Cloud listNebius, not stated$5.402026-11-09
H200Cloud listTogether AI (GPU Clusters, on-demand), not stated$5.992026-11-09
H200Cloud listCoreWeave, North America$6.302026-11-09
H200Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$7.912026-11-09
H200Cloud listFireworks (on-demand deployments), not stated; region-restricted deployments 1.5x$8.002026-11-09
H200Cloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$10.602026-11-09
H100NosanaNVIDIA H100 (premium)$1.5006:01:53
H100Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$1.992026-11-09
H100Akashnvidia h100 80Gi SXM5 (median of listings, lowest $2.55)$2.5606:02:01
H100Cloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$2.692026-11-09
H100Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$2.892026-11-09
H100Cloud listLambda, not stated$3.792026-11-09
H100Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$3.992026-11-09
H100Cloud listTogether AI (GPU Clusters, on-demand), not stated$3.992026-11-09
H100Cloud listDigitalOcean (GPU Droplets), not stated$4.412026-11-09
H100Cloud listNebius, not stated$4.502026-11-09
H100Cloud listLambda, not stated$4.592026-11-09
H100Cloud listLambda, not stated$4.892026-11-09
H100Cloud listPaperspace, not stated$5.952026-11-09
H100Cloud listCoreWeave, North America$6.162026-11-09
H100Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$6.882026-11-09
H100Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$6.882026-11-09
H100Cloud listFireworks (on-demand deployments), not stated; region-restricted deployments 1.5x$8.002026-11-09
H100Cloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$11.062026-11-09
H100Cloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$11.682026-11-09
H100Cloud listAzure, eastus (East US)$12.292026-11-09
A100 80GBCloud listRunPod (Community Cloud), not stated (pods across 30+ regions)$1.392026-11-09
A100 80GBAkashnvidia a100 80Gi SXM4 (median of listings, lowest $1.07)$1.4906:02:01
A100 80GBCloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$1.792026-11-09
A100 80GBCloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$1.792026-11-09
A100 80GBCloud listCoreWeave, North America$2.702026-11-09
A100 80GBCloud listLambda, not stated$3.192026-11-09
A100 80GBCloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$3.432026-11-09
A100 80GBCloud listAzure, eastus (East US)$3.672026-11-09
A100 80GBCloud listAzure, eastus (East US)$4.102026-11-09
A100 80GBCloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$5.072026-11-09
A100 40GBCloud listLambda, not stated$1.992026-11-09
A100 40GBCloud listLambda, not stated$1.992026-11-09
A100 40GBCloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$2.742026-11-09
A100 40GBCloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$3.672026-11-09
RTX PRO 6000NosanaNVIDIA Pro 6000 Server Edition (CC) (premium)$2.3406:01:53
L40SCloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$1.092026-11-09
L40SCloud listNebius, not stated$1.552026-11-09
L40SCloud listDigitalOcean (GPU Droplets), not stated$1.572026-11-09
L40SCloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$1.862026-11-09
L40SCloud listCoreWeave, North America$2.252026-11-09
RTX 6000 AdaCloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$0.9902026-11-09
RTX 6000 AdaCloud listDigitalOcean (GPU Droplets), not stated$1.572026-11-09
RTX A6000NosanaNVIDIA A6000 (premium)$0.40006:01:53
RTX A6000Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$0.5902026-11-09
RTX A6000Cloud listLambda, not stated$1.092026-11-09
RTX A6000Cloud listPaperspace, not stated$1.892026-11-09
RTX 5090Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$1.192026-11-09
RTX 4090NosanaNVIDIA 4090 (premium)$0.40006:01:53
RTX 4090Nosana4090 (Reserved) (premium)$0.77006:01:53
RTX 4090Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$0.8902026-11-09
RTX 3090Cloud listRunPod (Secure Cloud), not stated (pods across 30+ regions)$0.5002026-11-09
L4Cloud listGoogle Cloud, not labelled (selector default Iowa us-central1)$0.7072026-11-09
L4Cloud listAWS EC2, us-east-1 (US East N. Virginia), Linux$0.8052026-11-09
H100 NVLCloud listAzure, eastus (East US)$6.982026-11-09

Nosana: renter price including its network fee. io.net: listed price. Akash: the median listing, with the lowest beside it (Akash caches up to 5 minutes). Cloud list: on-demand list prices read 2026-11-09T17:32Z..17:44Z.

Held back

not shown above
  • RTX PRO 6000 on Nosana: $13.20/hthe offer is a machine of 6 GPUs, so its price is not per GPU-hour
  • B300 on Akash: $0.230/hunder 20% of the B300 live-market median (2.36/h across 2 prices)