Decodo vs Oxylabs: Which Residential Network Costs Less at Scale

On published list prices in September 2026, Decodo is cheaper than Oxylabs across the entire self-serve range, and the gap is not small. Under 125 GB a month it is frequently a 3x to 6x difference, because Oxylabs' residential plan ladder has very wide steps and Decodo's does not.

The flip happens at the top. Decodo's published ladder stops at 100 GB and $2.75/GB. Oxylabs publishes a 1 TB Corporate plan at $2.50/GB. On published numbers alone, Oxylabs holds the lower per-GB rate at terabyte scale, and Decodo at that volume becomes a sales conversation rather than a checkout page. Both negotiate above their published tiers, so past 1 TB you need two quotes, not two pricing pages.

Every Decodo figure below was read from Decodo's own residential proxy pricing page on 8 September 2026. Every Oxylabs figure was read from Oxylabs' own residential proxy pool page and public pricing page on the same date. Nothing here is a benchmark SparkProxy ran against either vendor, and no performance number below is a measurement we took. Where a vendor publishes a success rate or a response time, we label it as that vendor's claim. Tiers get renamed, promotional rates expire, and both companies discount at volume, so check both current pages before you spend anything on the strength of this table.

One naming note that trips up buyers reading older reviews: Decodo is the company formerly called Smartproxy, and it rebranded in 2025. A 2024 review of Smartproxy and a 2026 review of Oxylabs are two different price lists. Our Smartproxy alternatives and Oxylabs alternatives pages cover the wider field around both.

Decodo residential, as published on 8 September 2026. Prices are listed plus VAT where it applies.

Oxylabs residential, as published on 8 September 2026.

Read the two together and the shape is obvious. Decodo's entry rate of $3.75/GB undercuts Oxylabs' best self-serve rate of $4/GB, and Decodo's $4/GB pay as you go matches that Advanced tier with no commitment at all. Oxylabs only goes lower at the 1 TB Corporate plan.

Pricing pages quote a rate. You pay for a plan. The number that matters is the smallest published plan that covers your actual monthly consumption, and that is where the two ladders stop looking similar.

Two caveats. Decodo lists its prices plus VAT, so add that before comparing if it applies to you. And both vendors discount off-list at volume, so the 1 TB row is the start of a negotiation, not a verdict.

Here is the thing almost no comparison article surfaces: the cost of landing between tiers.

Oxylabs' ladder steps 5, 20, 125, 1000, a 6.25x jump from Basic to Advanced. If your real consumption is 30 GB a month, you cannot buy 30 GB. You buy the 125 GB Advanced plan at $500 and use under a quarter of it, which makes your effective rate about $16.67 per gigabyte actually consumed, not the $4 on the label. That is the single most common way buyers overpay for residential proxies.

Decodo's ladder steps 3, 10, 25, 50, 100, and the widest jump is 2.5x. The same 30 GB workload lands on the 50 GB plan at $150, an effective rate near $5 per gigabyte consumed. Still above the label, nothing like the overshoot.

The rule applies to every metered proxy vendor, not just these two. Compute your effective rate as plan price divided by gigabytes you will actually consume, not by gigabytes included. Then compare. A vendor with a higher headline rate and finer tiers routinely beats one with a lower headline rate and coarse tiers, and we work the same arithmetic through in understanding datacenter proxy pricing models.

Decodo's $4/GB pay as you go rate removes the tier-gap problem entirely for spiky or seasonal work. It costs more per gigabyte than any Decodo subscription tier, and it is still cheaper than overshooting a coarse tier by four times.

Both vendors bill bytes. So the largest variable in your invoice is not the rate you negotiated, it is how many bytes each useful record costs you to collect. That number is under your control, and it moves further than any price difference between these two companies.

Work the arithmetic on your own targets:

GB per month = (requests per month x average response bytes) / 1,073,741,824

Run 500,000 requests a month and the answer swings enormously depending on what you fetch:

Those page weights are illustrative inputs for the formula, not measurements of any site. Measure yours. The point is the spread: raw HTML against a fully rendered page is roughly eight times on that arithmetic, while Decodo's $2.75/GB against Oxylabs' $4/GB is 1.45 times. The engineering decision is worth several times the procurement decision.

Three changes usually cut a residential bill harder than switching vendors:

• Block images, fonts, media and analytics scripts in your headless browser. On image-heavy retail pages this alone often removes most of the transferred bytes.
• Hit the JSON endpoint the page calls instead of the page. Many product, search and pricing views are backed by an API returning a fraction of the payload.
• Send Accept-Encoding: gzip, br and confirm the target honours it. An uncompressed HTML response can be three to four times the compressed size, and you are billed on what crosses the wire.

Do that work before you sign anything. A team that halves its bytes per record and buys the more expensive vendor still pays less than a team that switches vendors and changes nothing.

Every figure in that table is the vendor's own marketing claim, including the pool sizes. Neither count is independently audited, and neither predicts what you will see on a hard target.

Two rows are worth acting on. Oxylabs lists coordinate-based targeting, which Decodo's published targeting list does not include. If your work needs a radius around a point rather than a city or ZIP, that capability difference can decide the purchase on its own. And Oxylabs publishes a specific sticky ceiling of 24 hours, while Decodo publishes sticky sessions without naming a maximum duration on that page. If you run long authenticated sessions, ask Decodo for the number in writing. The rest of the pre-purchase checklist is in what to evaluate when selecting a residential or datacenter proxy service.

The pool sizes themselves should carry very little weight. Past a few million addresses, no target's rate limiter can tell a 115M pool from a 175M one. What decides your success rate is whether the specific subnets you get routed onto are already burned on your target, and no published pool figure tells you that. Only a trial does. When to rotate and when to hold an IP is covered in rotating vs static residential proxies.

Buying friction is a real cost, and it is where the two differ most day to day.

Decodo publishes a 3-day free trial with 100 MB of residential traffic, plus a 14-day money-back guarantee that excludes free trial users. That is enough to run a few thousand lightweight requests against your real targets before paying anything, without talking to anyone.

Oxylabs states that a residential free trial is available once, on request through a contact form or by emailing support. That routes you through a person. For an enterprise buyer with a procurement process, that is often preferable, because an account manager gets attached early. For an engineer who wants to know by Thursday whether the IPs work on one target, it is a delay.

Budget for VAT too. Decodo lists prices plus VAT, either recoverable or a real addition to every figure in this article's Decodo column depending on where you are registered.

Buy Decodo when

• Your monthly consumption is under about 125 GB. On published list the gap in that range outweighs everything else in this comparison.
• Your volume is spiky or seasonal, and $4/GB pay as you go with no commitment beats guessing a tier.
• You want to test today without a sales call, using the 3-day, 100 MB trial.
• Your targeting needs stop at ZIP code or ASN level.

Buy Oxylabs when

• You need coordinate-based targeting, or a documented 24-hour sticky ceiling in writing.
• You are buying at 1 TB a month or more, where the published $2.50/GB Corporate rate sits below anything on Decodo's public ladder.
• Procurement requires enterprise contracting, account management and a long compliance paper trail. That is a legitimate reason to pay more, and it is often the real reason large buyers land here.
• You already run other Oxylabs products and want one contract and one invoice.

Neither list contains "has the bigger pool", because that criterion decides nothing at these sizes.

Residential IPs are the most expensive way to make a request, and plenty of teams buying them do not need them. Before comparing these two, confirm your target actually blocks datacenter ASNs. Many do not. Public catalogs, travel and pricing pages, and a lot of B2B collection run fine on datacenter IPs, and the full trade is in residential vs datacenter proxies.

If datacenter IPs work on your target, the billing unit changes and so does the arithmetic. SparkProxy sells rotating datacenter proxies on flat monthly plans with unlimited bandwidth and 30 days validity, priced by concurrent threads rather than by gigabyte:

Those plans draw on 1M+ datacenter IPs across 80+ countries, including more than 50,000 US datacenter IPs, reached through gateway.sparkproxy.io on port 11000 for HTTP and HTTPS, 11002 for sticky sessions and 13000 for SOCKS5. Each tier carries a speed cap under the fair usage policy, from 25 Mbps on Starter to 150 Mbps on Plus, and those are ceilings rather than guaranteed rates.

Put that next to the metered math. The 105 GB raw-HTML workload above costs $500 a month on Oxylabs Advanced at list, or roughly $295 on Decodo before VAT once you add pay as you go on top of the 100 GB plan. The same workload on a flat datacenter plan costs $240 whether it moves 105 GB or 400 GB. That comparison is only valid if your target accepts datacenter IPs. If it does not, no amount of flat pricing helps you.

If the reason you are shopping for residential IPs is anti-bot handling rather than geography, a managed scraping endpoint may cost less than either proxy bill, because failed retries stop being your problem. SparkProxy's Scraping API starts with 1,000 free credits and no card, then $49 for 250,000 credits a month at 50 concurrent requests, $99 for 1,000,000 at 100 concurrent, $249 for 3,000,000 at 200, and $599 for 8,000,000 at 400. A plain fetch costs 1 credit, a JavaScript render 5, a screenshot or PDF 10. That trade against running your own pool is laid out in web scraping API vs self-managed proxies.

Do not buy either vendor on a pricing page. Run the same test twice and compare one number.

1. Pick 1,000 real URLs from your actual target, not a test site. Include the hardest page type you scrape.
2. Run all 1,000 through each vendor's trial with identical code, headers and rendering settings. Change only the proxy endpoint.
3. Record three things per vendor: how many requests returned the data you wanted, total bytes transferred, and median latency.
4. Convert to gigabytes per 1,000 useful records. That is your real consumption unit.
5. Multiply by the tier you would actually land on, not the tier you would like to land on.

The arithmetic for step 5:

def cost_per_1k_records(plan_price, gb_included, gb_used, gb_per_1k_records, success_rate):
    """Effective cost per 1,000 USEFUL records on a metered proxy plan."""
    effective_rate = plan_price / max(gb_used, 0.001)   # what you really pay per GB
    return (gb_per_1k_records / success_rate) * effective_rate

# 30 GB/month workload, 0.9 GB per 1,000 useful records
print(cost_per_1k_records(150, 50, 30, 0.9, 0.94))   # finer tier: 30 GB used of 50
print(cost_per_1k_records(500, 125, 30, 0.9, 0.97))  # coarse tier: 30 GB used of 125

Feed in your own measured success rate, not either vendor's published one. A network three points better on success rate but four times the effective rate is not a better buy, which is exactly the situation the tier-gap section describes. For a fuller side by side of the flat model against one of these two, see SparkProxy vs Oxylabs.