Why Proxy Pricing Is So Confusing
You'd think buying proxies would be simple: pick a plan, pay, done. Instead you're staring at per-GB rates, per-IP/port pricing, per-request tiers, geo-premiums, and sticky-session surcharges — all from different providers who measure things differently. This guide breaks down every major proxy pricing model, shows you how to calculate your real monthly cost, and helps you decide which model fits your workload.
The single most important number in proxy pricing is your cost per successful request — not the headline rate on a provider's landing page. By the end of this article you'll be able to estimate that number for any use case.
The Three Proxy Billing Models
1. Per GB (Bandwidth-Based) — Residential & Mobile Dominant
Most residential and mobile proxy providers charge by the gigabyte of data transferred through their network. You pay for every byte that flows through the proxy — both request headers and response bodies.
- Typical range: $1–$15/GB depending on provider tier and geo.
- Best for: High-volume web scraping, SERP monitoring, e-commerce price checks — any workload where pages are small but you need diverse IPs.
- Watch out for: Bandwidth includes request headers and connection overhead, so your effective cost is always higher than a naive page-size × pages calculation.
2. Per IP / Per Port — Datacenter & ISP Dominant
Datacenter and static residential (ISP) proxies are usually sold by the IP address or by the port. You rent a fixed number of IPs for a month and get unlimited bandwidth.
- Typical range: $0.50–$5/IP/month for datacenter; $2–$15/IP/month for ISP proxies.
- Best for: Account management, long-running sessions, sneaker and ticketing bots, any workload that needs a stable identity.
- Watch out for: You pay whether or not you use the IP. If your bandwidth per IP is low, per-port pricing can be wasteful.
3. Per Request — SERP APIs & Specialized Services
A small number of providers — primarily SERP API services — charge per successful request regardless of payload size. This model abstracts away the proxy entirely.
- Typical range: $0.001–$0.01/request depending on volume and search engine.
- Best for: SERP tracking, lightweight API polling, teams that don't want to manage proxy infrastructure.
- Watch out for: Per-request pricing gets expensive fast if you're scraping large pages. A 500 KB product page at $0.01/request is orders of magnitude pricier than the same traffic on per-GB billing.
Calculating Real Cost Per Use Case
Let's walk through a concrete example. Suppose you're scraping 10 million product pages per month from an e-commerce site. Average page weight is 50 KB of HTML + 30 KB of headers/overhead = 80 KB per request.
Total bandwidth = 10,000,000 requests × 80 KB
= 800,000,000 KB
= ~763 GB/month
At a residential proxy rate of $3/GB:
Monthly cost = 763 GB × $3 = $2,289
Cost per successful request = $2,289 / 10,000,000 = $0.00023
Now compare that to per-request pricing at $0.005/request:
Monthly cost = 10,000,000 × $0.005 = $50,000
The per-request model is 22× more expensive for this use case. That's why bandwidth-based pricing dominates residential proxies: it scales with actual data transferred, not with an arbitrary request count.
Here's a quick calculator formula you can adapt:
BW_per_request_KB = avg_page_KB + overhead_KB # ~20-30 KB overhead
Total_GB = (requests_per_month × BW_per_request_KB) / 1_048_576
Monthly_cost = Total_GB × price_per_GB
Cost_per_request = Monthly_cost / requests_per_month
Hidden Costs That Blow Up Your Budget
Sticky-Session Upcharges
Some providers charge extra — either a percentage or a flat per-session fee — when you request a sticky session (an IP held for 10–30 minutes instead of rotating per request). If your scraper needs login sessions or multi-step workflows, budget an extra 10–25% on top of base bandwidth.
Geo-Premium Pricing
Not all countries cost the same. US, UK, German, and Australian IPs typically command a 20–50% premium over Southeast Asian, African, or Latin American IPs because residential supply is tighter and more expensive to source.
| Region | Typical Premium vs. Base Rate | Why It Costs More |
|---|---|---|
| US, UK, DE, AU | +20–50% | Higher demand, stricter ISP policies, lower peer supply |
| FR, IT, ES, CA | +10–25% | Moderate demand, decent supply |
| SEA (TH, VN, PH) | Base rate or –10% | Lower demand, abundant peer supply |
| LATAM (BR, MX) | Base rate | Moderate supply, growing demand |
Per-Request Overhead
Every HTTP request through a residential proxy carries overhead: DNS resolution, TLS handshake, TCP setup. This adds roughly 5–30 KB per request even if the response body is tiny. If you're making millions of API calls with small JSON responses, overhead can account for 40–60% of your total bandwidth.
Connection Failures & Retries
A request that times out still consumed bandwidth on the way in. Budget a 5–15% failure rate into your cost model depending on target site difficulty. A 10% failure rate means you're paying for 1.1× the bandwidth you actually need.
Budget-Tier vs. Premium-Tier Providers
Why does Provider A charge $1.50/GB while Provider B charges $8/GB? Here's where the extra money goes:
| Factor | Budget Tier ($1–3/GB) | Premium Tier ($5–15/GB) |
|---|---|---|
| Subnet diversity | Limited — many IPs from a few subnets | Broad — IPs spread across thousands of /24 blocks |
| ASN diversity | A handful of ISPs per country | Hundreds of ISPs, including regional carriers |
| Success rate | 85–92% on moderate targets | 95–99% on hard targets |
| Ops quality | Best-effort, community support | Dedicated account managers, priority routing |
| IP freshness | Shared pool, higher block rate | Frequently refreshed, lower block rate |
| Geo-coverage | 50–100 countries | 190+ countries with city-level targeting |
When budget-tier makes sense: You're scraping low-difficulty targets at high volume, and a 10% success-rate hit is tolerable. If your cost model works at $2/GB with retries, don't overspend on premium.
When premium-tier pays for itself: You're hitting aggressive anti-bot systems (sneaker sites, ticketing platforms, social media). The higher success rate means fewer retries, which means less wasted bandwidth. A $10/GB provider with 98% success can be cheaper per successful request than a $2/GB provider with 80% success.
# Effective cost comparison
Budget: $2/GB ÷ 0.80 success = $2.50/GB effective
Premium: $10/GB ÷ 0.98 success = $10.20/GB effective
# But factor in retry bandwidth:
Budget: $2/GB × 1.25 (retries) ÷ 0.80 = $3.13/GB effective
Premium: $10/GB × 1.02 (retries) ÷ 0.98 = $10.41/GB effective
# For easy targets (95%+ success on both):
Budget: $2/GB × 1.05 ÷ 0.95 = $2.21/GB effective
Premium: $10/GB × 1.01 ÷ 0.99 = $10.20/GB effective
For easy targets, budget-tier wins decisively. For hard targets, factor in the cost of failed requests and engineering time spent working around blocks.
Negotiating: When and How to Get Custom Pricing
Published rates are starting points. Here's when and how to negotiate:
Volume Thresholds
Most providers offer custom pricing once you're committing to 5 TB+/month on residential or 100+ IPs on datacenter. If you're at that scale, reach out — you can typically get 15–30% off published rates.
Commitment Discounts
Monthly plans are the most expensive. Providers will discount for:
- 3-month prepay: 5–10% off
- Annual commitment: 15–25% off
- Usage minimums: e.g., commit to 10 TB/month for a lower per-GB rate
SLA Tiers
Enterprise contracts often include SLA guarantees — uptime commitments, dedicated support channels, and sometimes guaranteed success rates. These add cost but reduce operational risk for production pipelines.
Negotiation Tips
- Come with real usage data: monthly GB, target sites, required success rates.
- Get quotes from 2–3 providers before negotiating — leverage is everything.
- Ask about overage rates. A low per-GB rate with punitive overage fees can be worse than a flat rate.
- Request a trial or proof-of-concept period before committing.
When to Switch Pricing Models
Your ideal pricing model depends on bandwidth per IP. Here's the decision framework:
| Scenario | Bandwidth/IP/month | Best Model | Why |
|---|---|---|---|
| Light scraping (SERPs, APIs) | < 1 GB/IP | Per GB (residential) | Low bandwidth per IP makes per-IP wasteful |
| Heavy scraping (product pages) | 1–10 GB/IP | Per GB (residential) | Rotation gives IP diversity; per-GB scales naturally |
| Account management | 10–100 GB/IP | Per IP (ISP/datacenter) | Stable identity + high bandwidth = per-IP is cheaper |
| Streaming / large downloads | > 100 GB/IP | Per IP (datacenter) | Unlimited bandwidth on per-IP plans wins decisively |
The crossover point is roughly 5–10 GB per IP per month. Below that, per-GB residential pricing is more efficient. Above that, per-IP datacenter or ISP proxies with unlimited bandwidth become cheaper.
Example: When Per-IP Wins
Say you manage 50 social media accounts, each needing a dedicated IP that uses ~30 GB/month of bandwidth.
Per-GB residential:
50 IPs × 30 GB × $3/GB = $4,500/month
Per-IP ISP proxy:
50 IPs × $8/IP/month = $400/month
Savings: $4,100/month (91% cheaper)
The math flips dramatically once bandwidth-per-IP crosses a threshold. Always model both before committing.
ProxyHat Pricing & Configuration Examples
ProxyHat offers residential, mobile, and datacenter proxies with transparent per-GB and per-IP pricing. Here's how to get started:
Residential Proxy (Per-GB)
# curl example — rotating residential proxy
curl -x http://user-country-US:PASSWORD@gate.proxyhat.com:8080 https://example.com
# Sticky session (30 min)
curl -x http://user-session-abc123-country-US:PASSWORD@gate.proxyhat.com:8080 https://example.com
# Python requests example
import requests
proxy = "http://user-country-US:PASSWORD@gate.proxyhat.com:8080"
response = requests.get("https://example.com", proxies={"http": proxy, "https": proxy})
print(response.status_code)
Datacenter Proxy (Per-IP)
# Node.js axios example — static datacenter IP
const axios = require("axios");
const proxy = "http://USERNAME:PASSWORD@gate.proxyhat.com:8080";
const res = await axios.get("https://example.com", {
proxy: { host: "gate.proxyhat.com", port: 8080, auth: { username: "USERNAME", password: "PASSWORD" } }
});
console.log(res.status);
Check ProxyHat pricing for current rates across residential, mobile, and datacenter tiers.
When NOT to Use Each Model
Honesty matters. Here are the scenarios where each model breaks down:
- Don't use per-GB residential if you need stable IPs for multi-step sessions and your bandwidth-per-IP is above 10 GB/month — you're overpaying.
- Don't use per-IP datacenter if the target site blocks datacenter ASNs — no amount of savings helps if your requests never succeed.
- Don't use per-request SERP APIs if you're scraping anything other than search engine result pages — the per-request cost is designed for tiny SERP responses, not 500 KB product pages.
- Don't use mobile proxies for high-volume scraping unless you specifically need mobile fingerprints — mobile bandwidth costs 3–5× more than residential.
Key Takeaways
1. Always model cost per successful request — not headline per-GB or per-IP rates. Factor in success rates, retries, and overhead.
2. The bandwidth-per-IP threshold decides your model — below ~5 GB/IP/month, go per-GB residential; above it, go per-IP datacenter/ISP.
3. Hidden costs add up fast — geo-premiums, sticky-session surcharges, and per-request overhead can increase your effective cost by 20–50%.
4. Premium providers can be cheaper per successful request on hard targets because fewer retries means less wasted bandwidth and engineering time.
5. Negotiate once you hit 5 TB/month — most providers offer 15–30% discounts for volume commitments.
Conclusion
Proxy pricing doesn't have to be a black box. Start by estimating your monthly bandwidth and requests, calculate your effective cost per successful request across models, and factor in the hidden costs we covered. If you're ready to test a provider with transparent pricing and strong subnet diversity, explore ProxyHat's plans — residential proxies start with flexible per-GB billing and datacenter IPs are available on straightforward per-IP terms.
For more on choosing the right proxy type for your workload, see our guides on residential vs datacenter proxies and web scraping use cases.
