GPU & inference providers
Where to rent GPUs or buy inference, with links to each provider’s own pricing page and any free credits we’ve verified for it. GPU rental prices change daily (see what the price data shows), so we link to the provider’s current prices instead of copying numbers here.
Disclosure: Perq may earn a referral fee from some providers on this page; any such link is labelled “referral link”. Listings are ranked by whether the provider has verified free credits on Perq, then alphabetically. Fees never affect whether a provider is listed or where. Operated by Avatar 8 LLC (ops@avatar33.com).
| Provider | Type | What it is | Free credits on Perq | Links |
|---|---|---|---|---|
| Baseten | Inference platform | Model deployment and inference billed for the time a model is deploying, scaling or serving. source | Pricing | |
| Cerebras | Inference API | Fast inference API with free, developer and enterprise plans. source | Pricing | |
| DigitalOcean (incl. Paperspace) | GPU cloud | GPU Droplets alongside its general developer cloud; Paperspace is part of DigitalOcean. source | Pricing | |
| Fireworks AI | Inference API | Inference platform for generative AI models; plans listed on its pricing page. source | Pricing | |
| Groq | Inference API | Inference cloud focused on fast inference. source | Pricing | |
| Lambda | GPU cloud | On-demand GPU instances, 1-Click Clusters and reserved capacity; rents H100 and B200 by the hour. source | Pricing | |
| Modal | Serverless GPU | Serverless GPUs and CPUs billed by the second, with nothing charged for idle time. source | Pricing | |
| Nebius | GPU cloud | AI cloud with on-demand NVIDIA GPUs from L40S and H100 up to B300 and GB300. source | Pricing | |
| Novita AI | Inference API + GPUs | 200+ model APIs, GPU instances and agent sandboxes. source | Pricing | |
| OpenRouter | Inference API | One API for hundreds of models from many providers, with per-model pricing and context length compared side by side. source | Pricing | |
| RunPod | GPU cloud | On-demand GPU Pods, Serverless GPU endpoints, Clusters and network storage, billed per second. source | Pricing | |
| Together AI | Inference API + GPU clusters | Serverless inference, dedicated endpoints, fine-tuning and GPU clusters. source | Pricing | |
| CoreWeave | GPU cloud | AI cloud with GPU, CPU and storage pricing published per instance. source | None verified | Pricing |
| DeepInfra | Inference API + GPUs | Pay-as-you-go model inference and GPU infrastructure with no long-term contracts. source | None verified | Pricing |
| Hyperbolic | GPU cloud + inference API | On-demand H100, H200 and B200 GPUs, reserved multi-node clusters and a serverless inference API. source | None verified | Pricing |
| Massed Compute | GPU cloud | On-demand GPU and CPU instances at published hourly rates, with no contracts or bandwidth charges. source | None verified | Pricing |
| NVIDIA DGX Cloud Lepton (formerly Lepton AI) | GPU marketplace | Connects developers to GPU compute across multiple cloud providers through one platform. source | None verified | Pricing |
| Prime Intellect | GPU cloud | Compute, training, inference and sandbox stack for training and deploying your own models. source | None verified | Pricing |
| SF Compute | GPU cloud | Buy and resell large-scale GPU clusters, managed from the terminal. source | None verified | Pricing Autoresearch referral link for SF Compute Autoresearch only |
| Shadeform | GPU marketplace | Access on-demand GPUs across multiple GPU clouds from one platform. source | None verified | Pricing |
| TensorDock | GPU marketplace | Cloud GPU marketplace for ML training, rendering and cloud gaming; it says it runs no referral program in order to keep margins thin. source | None verified | Pricing |
| Thunder Compute referral link | GPU cloud | Low-cost cloud GPUs for developers, billed by the minute with no commitments. source | None verified | Pricing Official site |
| Vast.ai referral link | GPU marketplace | Marketplace for renting GPUs; prices are set by supply and demand, with on-demand, interruptible and reserved rentals. source | None verified | Pricing Official site |
| Vultr | GPU cloud | Cloud provider with cloud GPU instances among its compute products. source | None verified | Pricing |
Provider descriptions paraphrase each provider’s own page (linked as “source”), checked Oct 23, 2026. Free credits come from the live Perq dataset and each has a verbatim source quote on its offer page. More: Cloud & GPU credits · LLM API credits · Compute crunch pack · GPU & inference pack.