Proxy Network Pricing Models Explained: Per GB, Per IP, Per Port and Per Request

Proxy pricing looks simple until a project goes live. A low per-GB rate can become expensive when pages are heavy, and a cheap per-IP plan can fail when targets block datacenter ranges. Providers listed in our proxy network software category, including Oxylabs, Smartproxy, SOAX and Webshare, use different billing units, which makes headline prices hard to compare.

This guide explains each pricing model, the hidden cost drivers and a formula that normalizes them. It is written for data, growth and engineering teams buying proxies for price monitoring, ad verification, market research and similar business use. Always collect web data lawfully, respecting site terms, applicable privacy law and robots exclusion signals; the EU GDPR applies when personal data is collected.

For provider selection steps, see our guide to choosing a proxy provider. Here we focus purely on money. Also review the OWASP Top 10 if your pipelines handle credentials or session tokens.

Per-GB Pricing

Per-GB pricing bills for the bandwidth consumed through the proxy, and it is the dominant model for residential and mobile proxies. You pay for every byte, including images, scripts and failed requests, so page weight, retries and headless browsers directly increase cost.

Per-GB plans suit workloads with small responses, such as API-like JSON endpoints. They suit poorly any workflow that renders full pages. Block images and fonts where legal and possible, and cache static assets to reduce bandwidth.

In the SaaSPodium directory, providers such as Geonode, IPRoyal and Rayobyte appear with entry prices that start low. Treat those figures as starting points: verify the billing unit, minimum commitment and per-GB rate at your expected volume.

Per-IP, Per-Port and Per-Thread Pricing

Per-IP pricing charges for dedicated or static addresses, usually datacenter or ISP proxies, often with generous or unmetered bandwidth. Per-port and per-thread models cap concurrency instead. They give predictable bills but can block quickly if targets flag the addresses.

• Dedicated IPs: stable reputation if only you use them, good for account-based tasks and allow-listing.
• Shared IPs: cheaper, but neighbors' behavior affects block rates.
• Concurrency limits: check threads or ports against your scheduler to avoid throttling.

Providers such as Proxy-Seller and Webshare lean toward individual proxy servers and per-IP plans. See our explanation of residential vs datacenter proxies for when each type works.

Per-Request and Scraping API Pricing

Per-request pricing bundles proxies, rotation, rendering and retries into a price per successful request or per thousand. It transfers anti-blocking effort to the vendor, which raises the unit price but lowers engineering time and makes budgeting simpler for teams without scraping specialists.

Several vendors, such as Oxylabs, which markets premium proxies and web scraping APIs, offer both raw proxies and managed APIs. Compare credit multipliers carefully: JavaScript rendering or premium geographies can consume several credits per call.

Ask how failed requests are counted. A plan that bills only successful responses aligns vendor and buyer incentives.

The Cost per Successful Request Formula

The fair way to compare plans is cost per successful request, which folds bandwidth, retries and block rates into one number. Estimate it from a one-day test with identical targets across vendors, rather than relying on advertised rates or success claims.

Success Rate is the share of requests returning usable content. Retry Rate is the average number of extra attempts per page. Average Page Size must be measured on the real target, rendered or not. A cheap plan with a 60 percent success rate often costs more than a premium plan at 95 percent.

Hidden Costs and Contract Terms

Look beyond list price. Minimum commitments, bandwidth expiry, geo-targeting surcharges, KYC requirements, and engineering time to manage blocks all change the true cost. Negotiate trial credits, rollover and clear refund terms before committing to annual contracts.

• Minimums and expiry: unused GB may not roll over.
• Targeting fees: city, ASN or carrier targeting is sometimes extra.
• Compliance checks: reputable providers screen use cases; expect questions about your intended targets.
• Engineering time: rotation, session logic and monitoring are real labor costs.

How to Run a Fair Pricing Test

Select 3 providers, run the same 10,000-request sample against your actual targets, log status codes, bandwidth and latency, then compute cost per successful request. Repeat at different times of day, and test the geographies you truly need before extending any commitment.

Include one premium network such as NetNut or SOAX, one mid-market option and one budget option so you see the spread. Read the vendor's official pricing page, for example Oxylabs, for the current billing units, and screenshot it for your records.

Finally, set alerts for bandwidth burn rate and success rate so cost surprises are caught in hours instead of at month end.

Frequently Asked Questions

How much do residential proxies cost?
Residential proxies are commonly billed per GB, and rates vary widely by provider, volume and commitment. Compare vendors on cost per successful request from a test, not on headline per-GB rates.

Is a scraping API cheaper than raw proxies?
Not on unit price, but often on total cost when you include engineering time, retries and anti-blocking maintenance. Model both options for your volume and team skills before deciding.

Can I change proxy providers without rewriting my scrapers?
Usually yes if you abstract the proxy endpoint and authentication in configuration. Test geotargeting parameters, session stickiness and error codes, which differ between providers.