Compute Prices
Compute Prices is a free price comparison platform for cloud GPU compute and LLM inference APIs. It aggregates pricing data from 85 providers, covering 71 GPU providers, 74 GPU models, 31 LLM providers, and 358 LLM models. The site is updated daily, with prices refreshed hourly, and provides detailed breakdowns for popular GPUs like H100, A100, B200, and popular LLMs such as GPT-OSS and DeepSeek models. Beyond the website, Compute Prices offers a free JSON API designed for AI agents and developers. With a free API key, users get 750 requests per day, week-over-week price movers, and 7-day price trends; a Pro tier extends history to 90 days. The API uses a stable shape across providers, eliminating the need for per-vendor scraping. The platform aims to help teams find the best rates for training, fine-tuning, and serving AI workloads. It also provides educational content, a weekly price report, and use-case-based GPU comparisons to support infrastructure decisions.
- Category
- LLM
- Pricing
- Free
Problem solved
Compute Prices solves the problem of fragmented and opaque pricing across cloud GPU and LLM API providers. It aggregates current prices into one continuously updated comparison site and API, enabling users to quickly find the most cost-effective provider for their AI compute needs without manually checking each vendor.
Use cases
- Compare on-demand and reserved GPU instance prices across cloud providers
- Find the cheapest API endpoint for serving LLM inference
- Monitor price fluctuations for AI infrastructure over time
- Retrieve structured GPU pricing data programmatically via the API for cost optimization
- Identify providers that meet budget constraints for training/fine-tuning jobs
- Evaluate regional or instance-type price variations
- Track weekly price movements to time infrastructure purchases
Audience
- Developers
- ML Engineers
- Data Scientists
- Startups
- Enterprise
Pros
- Covers 85 providers and 430+ models in one place
- Free API with 750 daily requests
- Daily updates and hourly price refreshes
- Unified JSON schema across providers
- Includes price trends and week-over-week movers
Cons
- Free tier limits historical trends to 7 days
- Limited to tracked providers; some regional or niche vendors may be missing
- API rate limit may be insufficient for high-volume production use
- No code samples for all major SDKs; primarily REST API
Official website: https://computeprices.com/
Canonical LambdaBase page: https://www.lambdabase.com/tools/compute-prices