Choosing a pricing model is about paying for the unit that matches how you actually run traffic. For the bigger context of networks, gateways, and billing units, see our proxy services handbook, then use this page to pick between per-IP, per-GB, subscriptions, or pay-as-you-go.
How Proxy Pricing Works
Pricing models differ by what you buy as the primary unit: an IP address, a gigabyte, a bundle, or a prepaid balance. Networks and workloads favor different units, so the “cheapest” line item is not always the lowest total cost.
In practice you pay for either 1) access to specific IPs or ports, 2) transferred traffic, 3) a tier with inclusions, or 4) a prepaid wallet. Datacenter and ISP pools often sell per IP or in bundles; residential and mobile more often meter by traffic. Replacement rules and limits ultimately decide your bill.
Quick model map
| Model | Billing unit | Typical networks | Strengths | Watch for |
| Per-IP | IPs or ports per month | Datacenter, ISP/static | Predictable identity, long sessions | Swap fees, minimum block sizes |
| Per-GB | Traffic volume (GB) | Residential, Mobile, rotating DC | Fine-grained scaling | Counting rules, overage, rounding |
| Subscriptions | Tiered bundles of IPs/GB/features | Any | Simple planning, discounts | Auto-renew, rollover, seat limits |
| Pay-as-you-go | Prepaid credit; charges by use | Any, trialing or bursty loads | Flexibility, no commitment | Higher unit price, balance expiry |
Per-IP Pricing: When a Static IP Makes Sense
Per-IP pricing rents specific addresses or ports, usually monthly, which fits steady workloads that need a stable egress identity. See details in Per-IP pricing guide.
Teams choose per-IP for allowlists and integrations that bind identity by source address. You pay for the slot whether you use it or not, so underutilization raises unit cost.
How billing works
You buy N IPs or ports, billed per period, sometimes grouped by subnet or location. Vendors may offer “Virgin” ranges, minimum blocks, or paid swaps if an IP is burned.
Check whether you are paying for exclusive IPs, shared endpoints, or a Gateway that maps to rotating backend IPs while keeping your port constant.
Where it fits best
Long-lived sessions, allowlist-only APIs, and jobs where reputation continuity beats raw scale. Per-IP also suits cases where you must pin CIDR ranges for firewall rules.
Pitfalls to verify
Swap quotas and fees, subnet variety, routing ASNs, and connection/thread caps per IP. Confirm IPv6 availability if you plan to mix stacks, and ask how QUIC/UDP is handled if your targets require it.
Per-GB (Bandwidth) Pricing: Counting and Overage
Per-GB pricing charges for transferred data and aligns cost with actual traffic. It is common for rotating residential and mobile pools where you do not pick specific IPs. See Per-GB pricing guide.
Success depends on how traffic is counted. Clarify what directions are billed, rounding granularity, and whether protocol overhead is included.
Counting rules that matter
Ask what counts: egress only, ingress plus egress, or both through the Gateway. Confirm rounding increments (per MB vs per GB) and how retries or 429 loops affect totals.
Overage and auto top-ups
Plans may throttle at quota, hard-stop, or auto-charge the next block. Decide whether automatic top-ups are acceptable and set alerts to avoid surprise bills.
When to choose Per-GB
Bursty, experimental, or wide-geo jobs that benefit from a large IP Pool without committing to fixed addresses. It also suits workloads where sticky windows are short and identity persistence is not critical.
Subscriptions and Bundles: What’s Included
Subscriptions package IP counts, GB, concurrency, and features into tiers, trading fine-grained control for simpler planning and a lower blended rate. See Subscription plans overview.
The key is to read what is actually included. Two plans with the same price can differ by targeting depth, allowed protocols, or the number of credentials.
Typical inclusions
A tier might bundle: a fixed number of IPs or ports, a GB allowance, country targeting, sticky sessions, and a concurrency cap. City/ASN/carrier targeting often sits in higher tiers.
Renewal and rollover rules
Check renewal timing, grace periods, and whether unused GB roll over. Some plans pause access at renewal if payment fails, others downshift to a free tier.
Choosing a tier
Pick by concurrency and targeting first, not headline IP or GB numbers. If you keep hitting stickiness or thread limits, a higher tier may be cheaper than over-optimizing usage.
Pay-As-You-Go: Flexibility vs Unit Cost
PAYG is prepaid credit with metered deductions for IPs, traffic, or time, ideal for pilots, intermittent tasks, or sudden peaks. See PAYG model details.
You pay a premium per unit for the right to stop anytime. Watch for balance expiry and minimum purchase blocks.
How providers implement PAYG
Common flows are wallet top-ups, one-time packs, or per-request billing behind an API Gateway. Some vendors require a small recurring fee to keep access to premium pools.
Avoiding run-outs
Enable alerts at multiple thresholds and keep a backup Gateway or a small static pool as a failover. Verify refund rules for unused credit and ask whether credit can fund add-ons like extra concurrency.
When PAYG is the right tool
Exploration, proof-of-concepts, event-driven spikes, and short campaigns where a subscription would sit idle the rest of the month.
Hidden Costs That Inflate Your Bill
Your invoice is shaped by limits as much as by unit price. Concurrency, stickiness, and targeting depth quietly change how many requests and retries you need to finish a job.
Check these before buying:
- Concurrency and thread caps per credential or per IP
- Sticky-session windows on rotating pools
- Targeting depth beyond country (city, ASN/ISP, carrier)
- Seats and auth methods (Username/Password seats, IP allowlists)
- Swap and replacement fees for per-IP plans
- Auto-renew and overage rules
- Protocol support (HTTP vs SOCKS, TLS versions, UDP/QUIC)
- Tax and currency (VAT, GST, FX fees)
Which Model Fits Each Network Type?
There is no universal mapping, but certain pairings are common because of how pools are sourced and operated. Treat the table below as a starting point and verify specifics with your provider.
| Network type | Common pricing | Why it’s common | Notes |
| Datacenter (DC) | Per-IP, subscriptions | Stable racks and static routing suit per-IP and bundled tiers | Rotating DC exists but often metered or per-port |
| ISP / Static Residential | Per-IP, subscriptions | Static consumer-labeled IPs used as stable endpoints | Swap rules and subnet diversity matter |
| Residential rotating | Per-GB, PAYG | Large IP Pool with session stickiness and time-based rotation | Counting rules and sticky TTL dominate cost |
| Mobile | Per-GB, per-device tiers, PAYG | Limited radio resources and carrier targeting drive metering | Confirm concurrent device caps and cooldowns |
How to Choose: A Practical Checklist
Pick the unit that matches your workload’s shape, then price the limits that drive retries. A short test against your actual targets beats theoretical comparisons.
- Define the session model (static/sticky vs rotating).
- Measure concurrency needs (threads per worker; per-credential and per-IP caps).
- List targeting must-haves (city/ASN/carrier often in higher tiers).
- Decide on commitment (start flexible on PAYG or a small sub, then scale).
- Check replacement and rollover (per-IP swaps; per-GB rollover; renewal timing).
- Validate protocol fit (HTTP/SOCKS, TLS ciphers, IPv6, UDP/QUIC if needed).
- Run a pilot (success rate, p50/p95 latency, GB per completed unit of work).
- Set guardrails (budget alerts, overage cutoffs, auto-renew preferences).
FAQs
What is the main difference between Per-IP and Per-GB pricing?
Per-IP sells you specific addresses for a period, ideal for stable identity, while Per-GB charges for transferred traffic, ideal for rotating pools and uneven usage.
Why do two plans with the same price perform differently?
Limits like concurrency caps, sticky TTL, and targeting depth change how many retries you need, which changes real cost and throughput.
Are subscriptions always cheaper than PAYG?
Subscriptions usually lower unit rates but only pay off if you use the inclusions. If your usage is spiky or uncertain, PAYG can be cheaper overall despite a higher unit price.
How do overage policies affect cost?
Hard stops can stall jobs, while auto-top-ups can overshoot budget. Pick the policy that matches your risk tolerance and set alerts.
Do I need IPv6 support?
If your targets support IPv6, mixing stacks can improve reach. Confirm v6 availability and billing parity with v4.
Related in this topic