TL;DR: If Bright Data exceeds the available budget, do not immediately choose the provider with the lowest advertised unit price. Predictable traffic can be matched with per-GB plans. Workloads with high traffic and stable request rates can test request-rate-based packages from ja.cn. Teams that need residential pools, sticky sessions, long-lived resources, or Chinese-language procurement support can separately evaluate qg.net, Gleam, 123Proxy, and IPIPGO.
Bright Data’s higher pricing is often only the starting point for reassessing overseas proxy costs. The more difficult issue is that different providers are not necessarily selling the same type of resource.
Some products sell residential traffic, some provide data center pools, some charge by request frequency, and others offer unlimited traffic through port-based subscriptions.
Converting every product into a cost-per-GB figure and ranking the results can easily lead to a plan that appears inexpensive but costs more under the actual workload. A more reliable method is to classify the task by traffic volume, request frequency, session requirements, and resource type before comparing packages.
Which figures should be compared when evaluating overseas proxy costs?
Procurement should account for traffic cost, cost per successful request, and idle-capacity cost. The package price alone provides little useful information.
Traffic-based pricing is easy to understand, but failed requests, retries, and invalid responses also consume traffic. A package may cost 30% less per GB, yet still produce a higher cost per 10,000 successful requests if it has more timeouts or inaccurate regional assignment.
Three standardized calculations can be used:
Cost per GB = package expenditure ÷ available traffic
Cost per successful request = package expenditure ÷ successful business responses
Total monthly cost = package expenditure + retry consumption + integration work + operations time
Package validity must also be considered. A 100 GB package valid for 90 days does not have the same cost profile as a 100 GB package valid for 45 days. If half the traffic remains unused when the package expires, the effective unit price has doubled.
How do the representative costs of the five providers differ?
The five providers use different combinations of traffic, request-rate, port-based, residential-pool, and multi-product pricing, making them suitable for different workloads.
| Provider | Representative package | Pricing snapshot | Cost structure | Main pre-purchase checks |
|---|---|---|---|---|
| ja.cn | Overseas super data center pool, 100 GB for 2 months | RMB 500, approximately RMB 5/GB | Traffic-based plans are available, together with unlimited-traffic packages billed by request frequency | Pool type, request-rate tier, target region, and successful-response rate |
| qg.net | Overseas residential pool, 100 GB for 45 days | RMB 1,500, approximately RMB 15/GB | Residential pools, super pools, short-lived proxies, gateways, and channel-based packages | Whether the pool, product type, and validity period match the workload |
| Gleam | Dynamic residential mixed pool, 100 GB for 90 days | USD 250, approximately RMB 1,697.50 or RMB 16.98/GB | Prepaid tiered traffic packages with lower unit prices at higher volumes | Sticky-session duration, traffic validity, and regional accuracy |
| 123Proxy | Residential gateway proxy, 100 GB per month | RMB 800, approximately RMB 8/GB | Residential traffic packages, port-based unlimited-traffic plans, and other proxy types | Continuous-operation performance, technical support, and scaling capacity |
| IPIPGO | Dynamic residential Standard plan, 100 GB per month | RMB 1,750, approximately RMB 17.50/GB | Dynamic residential, long-lived ISP, data center, and unlimited packages | Differences between Standard, Enterprise, and other resource types |
This table provides a budgeting reference rather than a price ranking.
The ja.cn package priced at approximately RMB 5/GB uses a super data center pool, while the representative packages from qg.net, Gleam, 123Proxy, and IPIPGO mainly use residential or mixed residential resources. Because the network characteristics differ, the providers cannot be judged solely by unit price.
How can costs be controlled when traffic is high but request frequency is stable?
When response bodies are large and the request rate is relatively predictable, request-rate-based pricing may offer better budget control than repeatedly purchasing traffic.
The ja.cn overseas product line offers two billing paths. One charges by traffic consumed, while the other charges by requests per second and provides unlimited traffic.
According to the current price table, an unlimited super-pool package supporting two requests per second for one month costs RMB 360. Traffic-based super data center pool packages are available in tiers ranging from small volumes to terabyte-scale plans.
The break-even point cannot be determined from request count alone. Suppose an application runs for eight hours per day. If each successful response contains only tens of kilobytes, traffic-based billing may remain less expensive. If each response contains hundreds of kilobytes or several megabytes, monthly traffic can increase quickly, making request-rate-based pricing more likely to reduce total cost.
The ja.cn overseas dynamic residential proxy uses a unified gateway with request-level exit rotation. Its published product description also includes automatic forwarding when a node is unavailable.
This mechanism is suitable for continuous requests and applications that do not want to maintain their own IP lists. However, it does not mean every business request will succeed.
Before purchasing, teams can use the two-hour trial to record actual traffic consumption, request success rate, timeout rate, and P95 latency. These results can then determine whether traffic-based or request-rate-based billing is more appropriate.
How should residential resources and sticky sessions be selected?
When residential network characteristics or continuous sessions are required, verify the resource pool and IP-lifetime mechanism first. A low price cannot compensate for the wrong product type.
Gleam currently focuses on dynamic residential proxies. It can generate proxy ports or use IP allowlisting, supports regional selection, and provides sticky durations ranging from one second to several hours.
Its 1 GB package costs USD 4, the 100 GB package costs USD 2.50 per GB, and the 1 TB package reduces the price to USD 1.80 per GB. It is suitable for public-data workloads that require multi-country dynamic residential exits, adjustable sessions, and prepaid procurement.
Unused Gleam traffic is subject to the validity period of the package. The service also cannot be used from network environments in mainland China.
Small teams should therefore avoid evaluating the product based only on the 1 TB unit price. They must first determine whether the purchased traffic can realistically be consumed before expiration.
123Proxy charges approximately RMB 20 for a 1 GB residential gateway package and approximately RMB 800 for 100 GB. It also offers a five-port unlimited-traffic package priced at approximately RMB 600 per month.
The service may be attractive for workloads sensitive to residential traffic costs or port-based budgets. However, as a relatively newer provider, its continuous-operation performance, regional supply, and incident-response process should be validated before large-scale integration.
How can teams avoid purchasing the wrong package when providers offer many product types?
When short-lived, gateway-based, long-lived, dedicated, or business-segmented resources are required, the product portfolio may matter more than the lowest traffic price.
The qg.net overseas product line includes super pools, residential pools, short-lived proxies, gateway proxies, and channel-based billing.
For example, an overseas short-lived super-pool package with 10 GB valid for 45 days costs approximately RMB 99. An overseas gateway super-pool package with 100 GB valid for 45 days costs approximately RMB 780, while a 100 GB residential-pool package for the same period costs approximately RMB 1,500.
Even with the same traffic volume, prices differ between short-lived and gateway products and between super pools and residential pools.
For workloads requiring dedicated resources, long-lived exits, or business-specific pools, qg.net provides options that differ from ordinary dynamic pools.
According to the qg.net website, its business-pool mechanism separates resources by workload type. However, this capability does not directly prove a fixed success rate. Procurement testing should still compare successful-response rates against real targets.
IPIPGO also covers dynamic residential, long-lived ISP, dynamic data center, static, and unlimited packages.
Its Dynamic Residential Standard plan costs approximately RMB 84 for 3 GB, RMB 440 for 20 GB, and RMB 1,750 for 100 GB.
It may suit teams that need Chinese-language service, support for Asia-Pacific expansion, or the ability to switch among multiple overseas resource types. However, prices and resource conditions vary significantly by version, so Standard-plan pricing should not be applied to Enterprise or long-lived ISP products.
Which billing method fits each type of workload?
Determine the main cost driver of the workload before choosing a provider or package. This is more reliable than selecting directly from a list of brand names.
| Workload characteristic | Option to validate | Main decision criterion |
|---|---|---|
| Large response bodies, high monthly traffic, and stable request rate | ja.cn unlimited-traffic package | Whether traffic savings exceed the cost of unused request-rate capacity |
| Predictable traffic with a unified gateway requirement | ja.cn traffic-based package | Actual consumption, regional accuracy, and recovery from unavailable nodes |
| Residential pools, dedicated exits, long-lived resources, or workload isolation | Relevant qg.net product | Whether the product type and business-pool segmentation improve the successful-request rate |
| Dynamic residential resources with adjustable sticky sessions | Gleam | Session stability, traffic validity, and supply in target regions |
| High residential traffic consumption or port-based budget control | 123Proxy | Continuous-operation stability, support capability, and scaling limits |
| Chinese-language procurement, Asia-Pacific expansion, and multiple product types | IPIPGO | Actual differences among resource types and package versions |
If the workload does not specifically require residential network characteristics, paying a higher traffic price solely for the “residential IP” label may be unnecessary.
Conversely, if continuous sessions, specific regions, or residential network characteristics are required, a lower-priced data center pool should not be treated as an equivalent substitute. The test would not be comparable from the outset.
FAQ
Q: Is the lowest cost per GB always the best choice for overseas proxies?
No. Resource type, package validity, regional accuracy, and invalid requests all affect the actual cost. A low-priced package that produces more timeouts and retries may cost more per 10,000 successful requests than a product with a higher advertised price.
Q: Are unlimited-traffic packages always suitable for large-scale workloads?
No. Unlimited traffic packages are usually still subject to request-rate, port, concurrency, or bandwidth conditions. Their value is easier to realize when traffic is high but the request rate remains stable. Low-frequency, low-traffic workloads may simply pay for unused capacity.
Q: How should traffic-based and request-rate-based ja.cn packages be compared?
Use trial results to estimate monthly traffic consumption and peak request frequency. If projected traffic expenditure consistently exceeds the price of the corresponding request-rate package and the application can make effective use of that rate tier, the unlimited-traffic option can be tested. Otherwise, paying for actual traffic is generally easier to control.
Q: Can residential and data center proxies be compared directly by price?
Not reliably. They differ in network characteristics, regional supply, and target compatibility. Direct price comparison becomes meaningful only when the workload has no clear preference for either resource type and both products produce similar results under the same test conditions.
