# RentGPU

> Cloud GPU price comparison for AI/ML workloads. Hourly rental prices across major providers, normalized into a single canonical catalog. Sister site to MiningBoard.com (which covers crypto miners on the same data layer).

## Catalog summary

- 270 GPU models tracked
- 31 active rental providers
- 40 AI models cataloged with affiliate access providers
- 11 workload categories

## GPU models

- [AMD Radeon R9 380 4GB](https://rentgpu.org/gpus/radeon-r9-380-4gb) — consumer, 4GB VRAM
- [AMD Radeon R9 380X](https://rentgpu.org/gpus/amd-r9-380x) — consumer, 4GB VRAM
- [AMD Radeon R9 390](https://rentgpu.org/gpus/amd-r9-390) — consumer, 8GB VRAM
- [AMD Radeon R9 Fury](https://rentgpu.org/gpus/radeon-r9-fury) — consumer, 4GB VRAM
- [AMD Radeon R9 Fury Nano](https://rentgpu.org/gpus/amd-r9-fury-nano) — consumer, 4GB VRAM
- [AMD Radeon RX 460](https://rentgpu.org/gpus/amd-rx-460) — consumer, 4GB VRAM
- [AMD Radeon RX 460 4GB](https://rentgpu.org/gpus/amd-rx-460-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 470](https://rentgpu.org/gpus/amd-rx-470) — consumer, 8GB VRAM
- [AMD Radeon RX 470 4GB](https://rentgpu.org/gpus/amd-rx-470-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 470 8GB](https://rentgpu.org/gpus/amd-rx-470-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 480](https://rentgpu.org/gpus/amd-rx-480) — consumer, 8GB VRAM
- [AMD Radeon RX 480 8GB](https://rentgpu.org/gpus/amd-rx-480-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 550](https://rentgpu.org/gpus/amd-rx-550) — consumer, 4GB VRAM
- [AMD Radeon RX 5500](https://rentgpu.org/gpus/amd-rx-5500) — consumer, 4GB VRAM
- [AMD Radeon RX 5500 XT](https://rentgpu.org/gpus/amd-rx-5500-xt) — consumer, 8GB VRAM
- [AMD Radeon RX 5500 XT 4GB](https://rentgpu.org/gpus/amd-rx-5500-xt-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 5500 XT 8GB](https://rentgpu.org/gpus/amd-rx-5500-xt-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 550 4GB](https://rentgpu.org/gpus/amd-rx-550-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 560](https://rentgpu.org/gpus/amd-rx-560) — consumer, 4GB VRAM
- [AMD Radeon RX 5600](https://rentgpu.org/gpus/amd-rx-5600) — consumer, 6GB VRAM
- [AMD Radeon RX 5600 XT](https://rentgpu.org/gpus/amd-rx-5600-xt) — consumer, 6GB VRAM
- [AMD Radeon RX 5600 XT 6GB](https://rentgpu.org/gpus/amd-rx-5600-xt-6gb) — consumer, 6GB VRAM
- [AMD Radeon RX 570](https://rentgpu.org/gpus/amd-rx-570) — consumer, 4GB VRAM
- [AMD Radeon RX 5700](https://rentgpu.org/gpus/amd-rx-5700) — consumer, 8GB VRAM
- [AMD Radeon RX 5700 8GB](https://rentgpu.org/gpus/amd-rx-5700-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 5700 XT](https://rentgpu.org/gpus/amd-rx-5700-xt) — consumer, 8GB VRAM
- [AMD Radeon RX 5700 XT 8GB](https://rentgpu.org/gpus/amd-rx-5700-xt-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 570 16GB](https://rentgpu.org/gpus/amd-rx-570-16gb) — consumer, 16GB VRAM
- [AMD Radeon RX 570 4GB](https://rentgpu.org/gpus/amd-rx-570-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 570 8GB](https://rentgpu.org/gpus/amd-rx-570-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 580](https://rentgpu.org/gpus/amd-rx-580) — consumer, 8GB VRAM
- [AMD Radeon RX 580 4GB](https://rentgpu.org/gpus/amd-rx-580-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 580 8GB](https://rentgpu.org/gpus/amd-rx-580-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 590](https://rentgpu.org/gpus/amd-rx-590) — consumer, 8GB VRAM
- [AMD Radeon RX 590 8GB](https://rentgpu.org/gpus/amd-rx-590-8gb) — consumer, 8GB VRAM
- [AMD Radeon RX 590 GME](https://rentgpu.org/gpus/amd-rx-590-gme) — consumer, 8GB VRAM
- [AMD Radeon RX 6400](https://rentgpu.org/gpus/amd-rx-6400) — consumer, 4GB VRAM
- [AMD Radeon RX 6500 XT](https://rentgpu.org/gpus/amd-rx-6500-xt) — consumer, 4GB VRAM
- [AMD Radeon RX 6500 XT 4GB](https://rentgpu.org/gpus/amd-rx-6500-xt-4gb) — consumer, 4GB VRAM
- [AMD Radeon RX 6600](https://rentgpu.org/gpus/amd-rx-6600) — consumer, 8GB VRAM

## Providers

- [Vast.ai](https://rentgpu.org/providers/vast-ai) — marketplace
- [RunPod](https://rentgpu.org/providers/run-pod) — first_party_cloud
- [TensorDock](https://rentgpu.org/providers/tensordock) — marketplace
- [Salad Cloud](https://rentgpu.org/providers/salad) — decentralized
- [io.net](https://rentgpu.org/providers/io-net) — decentralized
- [Akash Network](https://rentgpu.org/providers/akash) — decentralized
- [Render Network](https://rentgpu.org/providers/render-network) — decentralized
- [Fluence](https://rentgpu.org/providers/fluence) — decentralized
- [Cudo Compute](https://rentgpu.org/providers/cudo-compute) — marketplace
- [Clore.ai](https://rentgpu.org/providers/clore-ai) — marketplace
- [Spheron Network](https://rentgpu.org/providers/spheron) — decentralized
- [Lambda Labs](https://rentgpu.org/providers/lambda-labs) — first_party_cloud
- [DeepInfra](https://rentgpu.org/providers/deepinfra) — first_party_cloud
- [Novita AI](https://rentgpu.org/providers/novita-ai) — first_party_cloud
- [Lium](https://rentgpu.org/providers/lium) — marketplace
- [Thunder Compute](https://rentgpu.org/providers/thunder-compute) — first_party_cloud
- [Cyfuture AI](https://rentgpu.org/providers/cyfuture-ai) — first_party_cloud
- [AceCloud](https://rentgpu.org/providers/acecloud) — first_party_cloud
- [Amazon Web Services](https://rentgpu.org/providers/aws) — hyperscaler
- [DigitalOcean](https://rentgpu.org/providers/digitalocean) — hyperscaler
- [Google Cloud](https://rentgpu.org/providers/gcp) — hyperscaler
- [Hyperstack](https://rentgpu.org/providers/hyperstack) — first_party_cloud
- [Latitude.sh](https://rentgpu.org/providers/latitude) — first_party_cloud
- [Microsoft Azure](https://rentgpu.org/providers/azure) — hyperscaler
- [Nosana](https://rentgpu.org/providers/nosana) — marketplace
- [Oracle Cloud](https://rentgpu.org/providers/oracle_cloud) — hyperscaler
- [RunCrate](https://rentgpu.org/providers/runcrate) — first_party_cloud
- [Sesterce](https://rentgpu.org/providers/sesterce) — first_party_cloud
- [TensorWave](https://rentgpu.org/providers/tensorwave) — first_party_cloud
- [Theta EdgeCloud](https://rentgpu.org/providers/theta-edgecloud) — decentralized
- [Vultr](https://rentgpu.org/providers/vultr) — hyperscaler

## AI models

- [Google: Gemini 3.8 Flash](https://rentgpu.org/ai-models/gemini-3-8-flash) — by Google DeepMind, closed
- [Meta: Muse Spark 1.3](https://rentgpu.org/ai-models/muse-spark-1-3) — by Meta AI, open-weights
- [Anthropic: Claude Fable 5.1](https://rentgpu.org/ai-models/claude-fable-5-1) — by Anthropic, closed
- [GLM 5.3 Flash](https://rentgpu.org/ai-models/glm-5p3-flash) — by zai-org, open-weights
- [GLM-5.3](https://rentgpu.org/ai-models/glm-5p3) — by zai-org, open-weights
- [Qwen: Qwen3.8 Flash](https://rentgpu.org/ai-models/qwen3-8-flash) — by Alibaba (Qwen Team), open-weights
- [Z.ai: GLM 5.3 Flash](https://rentgpu.org/ai-models/glm-5-3-flash) — by Zhipu AI, open-weights
- [granite-4.2-30b](https://rentgpu.org/ai-models/granite-4-2-30b) — by IBM Research, open-weights
- [granite-4.2-3b](https://rentgpu.org/ai-models/granite-4-2-3b) — by IBM Research, open-weights
- [granite-4.2-8b](https://rentgpu.org/ai-models/granite-4-2-8b) — by IBM Research, open-weights
- [Tencent: Hy-MT2-7B](https://rentgpu.org/ai-models/hy-mt2-7b) — by Tencent, open-weights
- [Mistral: Ministral 8B](https://rentgpu.org/ai-models/ministral-8b) — by Mistral AI, open-weights
- [Ox Alpha](https://rentgpu.org/ai-models/ox-alpha) — by Stealth, closed
- [Tencent: Hy-MT2-1.8B](https://rentgpu.org/ai-models/hy-mt2-1-8b) — by Tencent, open-weights
- [Tencent: Hy-MT2-30B-A3B](https://rentgpu.org/ai-models/hy-mt2-30b-a3b) — by Tencent, open-weights
- [glm-5.3](https://rentgpu.org/ai-models/glm-5-3) — by Zhipu AI, open-weights
- [Qwen: Qwen3.8 27B](https://rentgpu.org/ai-models/qwen3-8-27b) — by Alibaba (Qwen Team), open-weights
- [Google: Gemini 3.7 Flash](https://rentgpu.org/ai-models/gemini-3-7-flash) — by Google DeepMind, closed
- [Qwen3.8-2.4T-A95B](https://rentgpu.org/ai-models/qwen3p8-2p4t-a95b) — by Alibaba (Qwen Team), open-weights
- [ByteDance Seed: Seed 2.1 Turbo](https://rentgpu.org/ai-models/seed-2-1-turbo) — by Bytedance Seed, open-weights
- [DeepSeek: DeepSeek V4 Pro 0813](https://rentgpu.org/ai-models/deepseek-v4-pro-0813) — by DeepSeek, open-weights
- [Qwen3.8-2.4T-A95B](https://rentgpu.org/ai-models/qwen3p8-max) — by Alibaba (Qwen Team), open-weights
- [Qwen: Qwen3.8 2.4T A95B](https://rentgpu.org/ai-models/qwen3-8-2-4t-a95b) — by Alibaba (Qwen Team), open-weights
- [SpaceXAI: Grok 4.6](https://rentgpu.org/ai-models/grok-4-6) — by xAI, closed
- [NVIDIA-Nemotron-3.5-Lightning](https://rentgpu.org/ai-models/nvidia-nemotron-3-5-lightning) — by Nvidia, open-weights
- [NVIDIA: Nemotron 3.5 Lightning](https://rentgpu.org/ai-models/nemotron-3-5-lightning) — by Nvidia, open-weights
- [Nemotron Lightning 3.5 30B A3B](https://rentgpu.org/ai-models/nemotron-lightning-3p5-30b-a3b) — by Nvidia, open-weights
- [Meta: Muse Glimmer 30B](https://rentgpu.org/ai-models/muse-glimmer-30b) — by Meta AI, open-weights
- [Sakana: Sakana Namazu](https://rentgpu.org/ai-models/sakana-namazu) — by Sakana, closed
- [Upstage: Solar Pro 4](https://rentgpu.org/ai-models/solar-pro4) — by Upstage, open-weights
- [Ling-3.0-flash](https://rentgpu.org/ai-models/ling-3-0-flash) — by Inclusionai, open-weights
- [Meta: Muse Spark 1.2](https://rentgpu.org/ai-models/muse-spark-1-2) — by Meta AI, open-weights
- [Qwen: Qwen3.8 Max](https://rentgpu.org/ai-models/qwen3-8-max) — by Alibaba (Qwen Team), open-weights
- [DeepSeek: DeepSeek V4 Flash 0731](https://rentgpu.org/ai-models/deepseek-v4-flash-0731) — by DeepSeek, open-weights
- [Thinking Machines: Inkling Small](https://rentgpu.org/ai-models/inkling-small) — by Thinkingmachines, closed
- [Qwen: Qwen3.7 Flash](https://rentgpu.org/ai-models/qwen3-7-flash) — by Alibaba (Qwen Team), open-weights
- [Claude Opus 5](https://rentgpu.org/ai-models/claude-opus-5) — by Anthropic, closed
- [Google: Gemini 3.5 Flash-Lite](https://rentgpu.org/ai-models/gemini-3-5-flash-lite) — by Google DeepMind, closed
- [Google: Gemini 3.6 Flash](https://rentgpu.org/ai-models/gemini-3-6-flash) — by Google DeepMind, closed
- [Meituan: LongCat 2.0](https://rentgpu.org/ai-models/longcat-2-0) — by Meituan, closed

## Use cases

- [Train large language models](https://rentgpu.org/use-cases/llm-training) — recommended 80GB+ VRAM
- [Fine-tune LLMs (LoRA / QLoRA)](https://rentgpu.org/use-cases/llm-fine-tuning) — recommended 24GB+ VRAM
- [Run LLMs (inference / serving)](https://rentgpu.org/use-cases/llm-inference) — recommended 12GB+ VRAM
- [Image generation (Stable Diffusion, Flux, …)](https://rentgpu.org/use-cases/diffusion) — recommended 8GB+ VRAM
- [Video generation](https://rentgpu.org/use-cases/video-gen) — recommended 24GB+ VRAM
- [Speech and audio AI](https://rentgpu.org/use-cases/speech-audio) — recommended 8GB+ VRAM
- [3D rendering and content creation](https://rentgpu.org/use-cases/3d-rendering) — recommended 12GB+ VRAM
- [General ML research and experiments](https://rentgpu.org/use-cases/ml-research) — recommended 12GB+ VRAM
- [Scientific compute (HPC, simulation)](https://rentgpu.org/use-cases/scientific-compute) — recommended 24GB+ VRAM
- [Text embeddings](https://rentgpu.org/use-cases/embedding) — recommended 2GB+ VRAM
- [Code generation](https://rentgpu.org/use-cases/code-generation) — recommended 16GB+ VRAM

## How prices are collected

We pull live listings from each provider's public API every hour, take the median hourly rate across active offers per (GPU model, provider) pair, and surface the lowest median. See https://rentgpu.org/methodology for the full methodology.