Amazon AWS alternatives
Compare 81 alternatives to Amazon AWS: providers that rent the same GPUs or serve the same models, ranked by how many they share, with live prices showing where each one is cheaper.
Amazon AWS's own prices →GPU cloud alternatives to Amazon AWS
67 providers rent at least one of the 10 GPUs Amazon AWS lists on demand, most shared first. Prices are each provider's cheapest on-demand rate per GPU-hour.
| Provider | Shared GPUs | Cheaper on | Biggest saving | Compare |
|---|---|---|---|---|
 Vast.ai | 10 of 10H100 SXM, H200, A100 SXM +7 more | 10 | B200 $7.50/hrvs $14.24/hr on Amazon AWS | Compare |
 Fluence Cloud | 9 of 10H100 SXM, H200, A100 SXM +6 more | 7 | H100 SXM $2.90/hrvs $6.88/hr on Amazon AWS | Compare |
 Modal | 9 of 10H100 SXM, H200, A100 SXM +6 more | 6 | HGX B300 $7.10/hrvs $17.80/hr on Amazon AWS | Compare |
 Omega Gradient | 8 of 10H200, A100 SXM, L40S +5 more | 7 | B200 $7.89/hrvs $14.24/hr on Amazon AWS | Compare |
 Runcrate | 8 of 10H100 SXM, H200, A100 SXM +5 more | 6 | B200 $7.80/hrvs $14.24/hr on Amazon AWS | Compare |
 Runpod | 8 of 10H100 SXM, H200, A100 SXM +5 more | 8 | HGX B300 $6.94/hrvs $17.80/hr on Amazon AWS | Compare |
 Aquanode | 7 of 10H100 SXM, H200, A100 SXM +4 more | 6 | B200 $9.35/hrvs $14.24/hr on Amazon AWS | Compare |
 Baseten | 7 of 10H100 SXM, H200, A100 SXM +4 more | 3 | B200 $9.98/hrvs $14.24/hr on Amazon AWS | Compare |
 gpu.ai | 7 of 10H100 SXM, H200, A100 SXM +4 more | 7 | HGX B300 $11.24/hrvs $17.80/hr on Amazon AWS | Compare |
 Hugging Face | 7 of 10H100 SXM, H200, A100 SXM +4 more | 6 | H200 $5.00/hrvs $7.91/hr on Amazon AWS | Compare |
 Oracle Cloud | 7 of 10H100 SXM, H200, A100 SXM +4 more | 2 | HGX B300 $15.00/hrvs $17.80/hr on Amazon AWS | Compare |
 Sesterce | 7 of 10H100 SXM, H200, A100 SXM +4 more | 5 | H100 SXM $2.75/hrvs $6.88/hr on Amazon AWS | Compare |
 Shadeform | 7 of 10H100 SXM, H200, A100 SXM +4 more | 5 | H100 SXM $2.50/hrvs $6.88/hr on Amazon AWS | Compare |
 Verda | 7 of 10H100 SXM, H200, A100 SXM +4 more | 6 | HGX B300 $11.57/hrvs $17.80/hr on Amazon AWS | Compare |
 Google Cloud | 6 of 10H100 SXM, H200, A100 SXM +3 more | 2 | Tesla T4 $0.350/hrvs $0.526/hr on Amazon AWS | Compare |
 IO.NET | 6 of 10H100 SXM, H200, A100 SXM +3 more | 6 | H100 SXM $2.60/hrvs $6.88/hr on Amazon AWS | Compare |
 Lium.io | 6 of 10H100 SXM, H200, A100 SXM +3 more | 6 | B200 $6.00/hrvs $14.24/hr on Amazon AWS | Compare |
 Lyceum | 6 of 10H100 SXM, H200, A100 SXM +3 more | 5 | HGX B300 $11.09/hrvs $17.80/hr on Amazon AWS | Compare |
 Spheron | 6 of 10H100 SXM, H200, A100 SXM +3 more | 6 | HGX B300 $11.52/hrvs $17.80/hr on Amazon AWS | Compare |
 Together AI | 6 of 10H100 SXM, H200, A100 SXM +3 more | 5 | HGX B300 $9.99/hrvs $17.80/hr on Amazon AWS | Compare |
 Civo | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | HGX B300 $7.49/hrvs $17.80/hr on Amazon AWS | Compare |
 CoreWeave | 5 of 10H100 SXM, H200, A100 SXM +2 more | 4 | B200 $8.60/hrvs $14.24/hr on Amazon AWS | Compare |
 Crusoe | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | HGX B300 $8.96/hrvs $17.80/hr on Amazon AWS | Compare |
 Deep Infra | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | HGX B300 $5.89/hrvs $17.80/hr on Amazon AWS | Compare |
 Gcore | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | HGX B300 $6.73/hrvs $17.80/hr on Amazon AWS | Compare |
 Hexgrid Cloud | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | B200 $7.50/hrvs $14.24/hr on Amazon AWS | Compare |
 Hyperstack | 5 of 10H100 SXM, H200, A100 SXM +2 more | 5 | HGX B300 $7.40/hrvs $17.80/hr on Amazon AWS | Compare |
 Lambda Labs | 5 of 10H100 SXM, A100 SXM, B200 +2 more | 4 | B200 $6.69/hrvs $14.24/hr on Amazon AWS | Compare |
 Massed Compute | 5 of 10H100 SXM, A100 SXM, L40S +2 more | 5 | HGX B300 $6.60/hrvs $17.80/hr on Amazon AWS | Compare |
 Microsoft Azure | 5 of 10H100 SXM, H200, A100 SXM +2 more | 1 | Tesla T4 $0.343/hrvs $0.526/hr on Amazon AWS | Compare |
 Nebius | 5 of 10H100 SXM, H200, L40S +2 more | 5 | HGX B300 $9.50/hrvs $17.80/hr on Amazon AWS | Compare |
 Seeweb | 5 of 10H100 SXM, H200, L40S +2 more | 5 | HGX B300 $5.60/hrvs $17.80/hr on Amazon AWS | Compare |
 UpCloud | 5 of 10H100 SXM, L40S, B200 +2 more | 5 | HGX B300 $5.85/hrvs $17.80/hr on Amazon AWS | Compare |
 VoltageGPU | 5 of 10H100 SXM, H200, L40S +2 more | 4 | HGX B300 $13.30/hrvs $17.80/hr on Amazon AWS | Compare |
 Akash Network | 4 of 10H100 SXM, A100 SXM, A10 +1 more | 4 | H100 SXM $2.04/hrvs $6.88/hr on Amazon AWS | Compare |
 DigitalOcean | 4 of 10H100 SXM, H200, L40S +1 more | 4 | HGX B300 $6.17/hrvs $17.80/hr on Amazon AWS | Compare |
 fal.ai | 4 of 10H100 SXM, H200, B200 +1 more | 4 | B200 $7.99/hrvs $14.24/hr on Amazon AWS | Compare |
 Jarvis Labs | 4 of 10H100 SXM, H200, A100 SXM +1 more | 4 | H100 SXM $3.49/hrvs $6.88/hr on Amazon AWS | Compare |
 OpenRelay | 4 of 10H100 SXM, B200, HGX B300 +1 more | 4 | HGX B300 $8.49/hrvs $17.80/hr on Amazon AWS | Compare |
 OVHcloud | 4 of 10H200, L40S, L4 +1 more | 3 | H200 $6.20/hrvs $7.91/hr on Amazon AWS | Compare |
 Scaleway | 4 of 10H100 SXM, L40S, HGX B300 +1 more | 3 | HGX B300 $8.42/hrvs $17.80/hr on Amazon AWS | Compare |
 Daytona | 3 of 10H100 SXM, H200, HGX B300 | 3 | HGX B300 $7.10/hrvs $17.80/hr on Amazon AWS | Compare |
 Denvr Dataworks | 3 of 10H100 SXM, A100 SXM, A100 SXM 40GB | 3 | H100 SXM $2.45/hrvs $6.88/hr on Amazon AWS | Compare |
 GMI Cloud | 3 of 10H100 SXM, H200, B200 | 3 | B200 $5.00/hrvs $14.24/hr on Amazon AWS | Compare |
 Hyperbolic | 3 of 10H100 SXM, H200, B200 | 3 | B200 $5.99/hrvs $14.24/hr on Amazon AWS | Compare |
 Koyeb | 3 of 10H100 SXM, A100 SXM, L40S | 3 | H100 SXM $2.50/hrvs $6.88/hr on Amazon AWS | Compare |
 Pacific Machines | 3 of 10H100 SXM, H200, B200 | 3 | B200 $6.14/hrvs $14.24/hr on Amazon AWS | Compare |
 Packet AI | 3 of 10A100 SXM, L40S, B200 | 3 | B200 $3.75/hrvs $14.24/hr on Amazon AWS | Compare |
 Atlas Cloud | 2 of 10H100 SXM, H200 | 2 | H200 $3.50/hrvs $7.91/hr on Amazon AWS | Compare |
 Cudo Compute | 2 of 10H100 SXM, L40S | 2 | H100 SXM $1.79/hrvs $6.88/hr on Amazon AWS | Compare |
 Hinode | 2 of 10L40S, L4 | 0 | Not cheaper | Compare |
 Latitude.sh | 2 of 10H100 SXM, HGX B300 | 2 | HGX B300 $8.00/hrvs $17.80/hr on Amazon AWS | Compare |
 Novita AI | 2 of 10H100 SXM, L40S | 2 | H100 SXM $3.39/hrvs $6.88/hr on Amazon AWS | Compare |
 Oblivus | 2 of 10H100 SXM, H200 | 2 | H200 $3.99/hrvs $7.91/hr on Amazon AWS | Compare |
 Paperspace | 2 of 10H100 SXM, A100 SXM | 2 | H100 SXM $5.95/hrvs $6.88/hr on Amazon AWS | Compare |
 ParalonCloud | 2 of 10H100 SXM, A100 SXM 40GB | 2 | H100 SXM $2.20/hrvs $6.88/hr on Amazon AWS | Compare |
 QuantaCloud | 2 of 10A100 SXM, L40S | 2 | A100 SXM $1.49/hrvs $3.43/hr on Amazon AWS | Compare |
 Theta EdgeCloud | 2 of 10H200, Tesla T4 | 2 | H200 $3.69/hrvs $7.91/hr on Amazon AWS | Compare |
 Thunder Compute | 2 of 10H100 SXM, A100 SXM | 2 | H100 SXM $3.20/hrvs $6.88/hr on Amazon AWS | Compare |
 AceCloud | 1 of 10L40S | 0 | Not cheaper | Compare |
 Bentaus | 1 of 10HGX B300 | 1 | HGX B300 $6.50/hrvs $17.80/hr on Amazon AWS | Compare |
 centron | 1 of 10A100 SXM | 1 | A100 SXM $2.00/hrvs $3.43/hr on Amazon AWS | Compare |
 CloudGPU | 1 of 10A100 SXM | 1 | A100 SXM $1.59/hrvs $3.43/hr on Amazon AWS | Compare |
 Orion AI Factory | 1 of 10B200 | 1 | B200 $9.99/hrvs $14.24/hr on Amazon AWS | Compare |
 Voltage Park | 1 of 10H100 SXM | 1 | H100 SXM $1.99/hrvs $6.88/hr on Amazon AWS | Compare |
 Vultr | 1 of 10L40S | 1 | L40S $1.67/hrvs $1.86/hr on Amazon AWS | Compare |
 Zettabyte | 1 of 10H100 SXM | 1 | H100 SXM $2.51/hrvs $6.88/hr on Amazon AWS | Compare |
LLM API alternatives to Amazon AWS
28 providers serve at least one of the 51 models Amazon AWS prices per token, most shared first. Input and output rates are compared separately on each shared model at standard pricing.
| Provider | Shared models | Lower input price on | Lower output price on | Compare |
|---|---|---|---|---|
 OpenRouter | 34 of 51DeepSeek V3.2, Claude Opus 4.6, GLM-4.7 +31 more | 17 | 15 | Compare |
 Prime Intellect | 18 of 51DeepSeek V3.2, Claude Opus 4.6, GLM-4.7 +15 more | 9 | 7 | Compare |
 RouterPlus | 13 of 51GPT-OSS-120B, GPT-OSS-20B, DeepSeek V3.2 +10 more | 8 | 8 | Compare |
| Novita AI | 12 of 51GPT-OSS-120B, GLM-4.7, GLM-5 +9 more | 5 | 6 | Compare |
| DigitalOcean | 11 of 51GPT-OSS-120B, GPT-OSS-20B, DeepSeek V3.2 +8 more | 7 | 6 | Compare |
| Deep Infra | 10 of 51GPT-OSS-120B, GPT-OSS-20B, DeepSeek V3.2 +7 more | 9 | 9 | Compare |
| IO.NET | 10 of 51GPT-OSS-120B, GPT-OSS-20B, DeepSeek V3.2 +7 more | 5 | 5 | Compare |
| Together AI | 10 of 51GPT-OSS-120B, GPT-OSS-20B, GLM-4.7 +7 more | 4 | 4 | Compare |
| GMI Cloud | 7 of 51GPT-OSS-120B, DeepSeek V3.2, GLM-4.7 +4 more | 3 | 4 | Compare |
 Anthropic | 4 of 51Claude Opus 4.6, Claude Haiku 4.5, Claude Sonnet 4.5 +1 more | 4 | 4 | Compare |
| OVHcloud | 4 of 51GPT-OSS-120B, GPT-OSS-20B, Mistral 7B +1 more | 4 | 4 | Compare |
 PopNow | 3 of 51Claude Opus 4.6, Claude Sonnet 4.5, Claude Opus 4.5 | 3 | 3 | Compare |
| Crusoe | 2 of 51GPT-OSS-120B, Nemotron 3 Nano 30B | 2 | 2 | Compare |
 EcoHash | 2 of 51GPT-OSS-20B, Llama 3.1 8B | 2 | 1 | Compare |
 Geodd | 2 of 51GPT-OSS-120B, GPT-OSS-20B | 2 | 2 | Compare |
 Groq | 2 of 51GPT-OSS-120B, GPT-OSS-20B | 0 | 0 | Compare |
| Hyperstack | 2 of 51GPT-OSS-120B, Llama 3.1 8B | 2 | 2 | Compare |
 IteraCompute | 2 of 51Kimi K2.5, MiniMax M2.5 | 2 | 2 | Compare |
| OpenRelay | 2 of 51GPT-OSS-120B, GPT-OSS-20B | 1 | 0 | Compare |
| Runcrate | 2 of 51Ministral 3 14B, Nemotron Nano 12B v2 VL | 0 | 0 | Compare |
| Baseten | 1 of 51GPT-OSS-120B | 1 | 1 | Compare |
 Cerebras | 1 of 51GPT-OSS-120B | 0 | 0 | Compare |
| CloudGPU | 1 of 51Claude Opus 4.6 | 1 | 1 | Compare |
 Empirio Labs | 1 of 51GLM 4.7 Flash | 0 | 0 | Compare |
 Heabsy | 1 of 51DeepSeek V3.2 | 1 | 1 | Compare |
 Mistral AI | 1 of 51Ministral 3 14B | 0 | 0 | Compare |
 SambaNova | 1 of 51GPT-OSS-120B | 0 | 1 | Compare |
| Scaleway | 1 of 51GPT-OSS-120B | 0 | 0 | Compare |
Frequently Asked Questions
What are the best alternatives to Amazon AWS for renting GPUs?
Ranked by how many of the same GPUs they rent on demand, the closest alternatives to Amazon AWS are Vast.ai (10 shared GPUs), Fluence Cloud (9 shared GPUs) and Modal (9 shared GPUs). Amazon AWS lists 10 GPUs on demand, and 67 other providers rent at least one of them.
Is there a cheaper alternative to Amazon AWS?
65 of the 67 providers renting the same GPUs list at least one of them for less on demand. The largest gap today is the HGX B300 on Seeweb, at $5.60 per GPU-hour against $17.80 on Amazon AWS.
Which providers serve the same models as Amazon AWS?
28 providers serve at least one of the 51 models Amazon AWS prices per token. The most overlap is at OpenRouter (34 models), Prime Intellect (18 models) and RouterPlus (13 models).
Where are the same models cheaper than on Amazon AWS?
21 providers charge less per input token than Amazon AWS on at least one shared model, and 21 providers charge less per output token. The two are counted separately because the cheapest host for input tokens is often not the cheapest for output.
How are these Amazon AWS alternatives chosen?
An alternative is any provider that currently prices at least one of the same GPUs on demand or the same models at standard per-token rates. They are ordered by how many they share, using the current prices we track for every provider. Sponsorship and paid placement have no effect on the order.