Launched October 27, 2026
CloudGPU

CloudGPU

Pay-as-you-go GPU rental and an OpenAI-compatible AI API, on one prepaid balance.

One time feeDevOpsAIAPI365 impressions#67 of its week3 comments

Comments

>log in to comment
  • yang gao[maker]· 24d ago

    Hi DevHunt! Maker of CloudGPU here. GPU rental by the minute plus an OpenAI-compatible AI API, on one prepaid balance. If you try it, tell me what confused you, that's what I'll fix first.

    • Fatisco Stack Industries· 15d ago

      If an instance takes two minutes to start and then fails to come up, are those minutes taken off your prepaid balance, or does a failed start cost nothing?

    • yang gao[maker]· 23h ago

      Sorry for the slow reply. If provisioning fails and the instance never starts, our published policy is to return the pre-authorised amount to your balance automatically, with no GPU charge. An instance that starts but is unusable is a different case: please send the instance ID to support@cloudgpu.app so we can check it. Policy: https://cloudgpu.app/refunds

CloudGPU combines GPU rental and an OpenAI-compatible model API on one prepaid balance. GPU instances support workloads such as fine-tuning, inference, ComfyUI, Ollama, LLaMA-Factory and JupyterLab. Billing depends on the selected configuration: per-minute options return unused prepaid minutes when destroyed; whole-hour options charge each started hour in full, with no refund for unused time on voluntary stop. The quote shows the billing period, minimum charge and initial prepayment before confirmation. Startup time and stock vary. The model API at https://cloudgpu.app/v1 offers access to models including DeepSeek, GLM, Kimi, MiniMax, FLUX and Claude. CloudGPU is an independent service, not affiliated with model vendors. This is prepaid usage, not a perpetual licence or a subscription. See https://cloudgpu.app/pricing for current rates and https://cloudgpu.app/docs for instance and Workspace storage rules.

More DevOps tools →80 free dev tools →

CloudGPU offers GPU rentals and an OpenAI-compatible model API on one prepaid balance. GPU billing is per minute or started hour, depending on the selected configuration.

for
AI developers and teams needing on-demand GPU compute or model inference without long-term contracts.
pricing
freemium

Key features

6 features of CloudGPU
  • Configuration-specific GPU billing — Per-minute options return unused prepaid minutes when destroyed. Whole-hour options charge each started hour in full. Check the billing period and minimum charge before confirming.
  • OpenAI-compatible API gateway — Access models such as DeepSeek, Claude, Flux and others via a standard OpenAI-compatible endpoint.
  • One-click model deployment — Deploy supported templates with a public endpoint where supported. Startup time varies and is not guaranteed.
  • Raw GPU access — SSH into the rented GPU for full control, install any software, and destroy the instance anytime.
  • Multiple payment methods — Top up via card, Alipay, Google/Apple Pay, USDT (TRC-20) or bank transfer; credits never expire.
  • Transparent pricing — Hourly rates listed for each GPU model and per-token rates for model APIs, with no hidden or egress fees.

Use cases

  • Fine-tune LLMs on rented GPUs with the billing period and minimum charge shown before confirmation
  • Run inference for image generation models like Flux via a simple API endpoint
  • Develop AI agents that need occasional GPU bursts without maintaining hardware

CloudGPU vs alternatives

CloudGPUOpenRouterReplicateluoluoAIVergora
Best forPay-as-you-go GPU compute & model APIsUnified model API accessModel execution as a serviceCheap multi-model APIMulti-model access without compute
PricingFreemiumSubscriptionSubscriptionOne time feeSubscription
DevHunt upvotes1005224
LaunchedOct 2026——Sep 2026Oct 2026

CloudGPU FAQ

How is GPU time billed?+

Per-minute configurations have a 1-minute minimum and return unused prepaid minutes on destruction. Whole-hour configurations have a 1-hour minimum and charge each started hour in full; unused time is not refunded on voluntary stop. The quote shows the billing period, minimum charge and initial prepayment.

How do I use the model API?+

Set the base URL to https://cloudgpu.app/v1 and provide your API key in any OpenAI-compatible SDK or client.

What payment methods are accepted?+

You can top up with Visa/Mastercard, Alipay, Google Pay, Apple Pay, USDT (TRC-20) or bank transfer for larger amounts.

Is there any commitment or minimum spend?+

There is no subscription. GPU time follows the selected configuration's per-minute or started-hour billing and minimum charge; model API usage is billed by tokens. See cloudgpu.app/pricing before confirming.

What happens to data on a destroyed instance?+

Destroying an instance erases its local disk. Eligible funded accounts can explicitly save project files in Workspace, subject to quotas and retention rules. Workspace is not an automatic disk backup. Save or download results before destroying an instance; see cloudgpu.app/docs.

Do credits expire?+

No, prepaid credits never expire and can be used for any GPU model or model API.

Edited by the makers, based on cloudgpu.app · Oct 25, 2026. Details may change; check the official site.