Browser API benchmarks

Sample tested 14 min ago· n = 1,284· Tokyo· hourly · fixed demo clockMethodology →

Test type

ProviderRegion

WorkloadTargets30D · 30 runs · 38,520 requests

Independent measurementDemo dataFetchGauge probes · self-funded accounts · no vendor-supplied dataMethodology v2.3 →

Last tested

14 min ago · hourly

Sample size

n = 1,284 / run · 924,480 / 30D

Region

Tokyo (ap-northeast-1)

Test type

Browser API · e-commerce workload

Configuration

JS off · 50 conc · 30s timeout · 3 retries

Targets

50 domains · top-traffic e-commerce

Simulated protocol · all figures and freshness timestamps are fixtures, not real measurements.

Median success rate

96.9%

+0.8 ptvs oldest stored run in 30D

Median p50 latency

1.80s

−40 msvs oldest stored run in 30D

Median p95 latency

4.60s

+0.46 svs oldest stored run in 30D

Median error rate

3.1%

−0.8 ptvs oldest stored run in 30D

Cost / 1k successful

$1.70

−$0.01vs oldest stored run in 30D

Reliability score

±00–100 · median of stored runs

Aug 11Aug 29Sep 8Sep 9Sep 10

p50 latencyseconds

p95 latencyseconds

Ranking · 30DClick a column to sort. Reliability = weighted success, p95 stability and error diversity.

#	Provider
01	BBright Data	98.4%	1.42s	3.90s	1.6%	$2.85	92	0.4%
02	ZZenRows	96.9%	1.80s	4.60s	3.1%	$1.70	86	0.9%
03	AApify	95.7%	2.10s	5.20s	4.3%	$1.25	81	1.2%

Demo dataLast tested Sep 10, 2026 11:46 UTCn = 1,284 / runTokyo30 runs shown

Sample: 38,520 requests · 3 retries max · timeout 30sTrend column: stored runs only · raw CSV

30D · by success

01Bright Data98.4%

98.4%1.42s$2.85 / 1k ok92 rel.

02ZenRows96.9%

96.9%1.80s$1.70 / 1k ok86 rel.

03Apify95.7%

95.7%2.10s$1.25 / 1k ok81 rel.

Tap a provider for pricing, capabilities and 30-day history.

Error breakdown share of failed requests

Provider	403	429	Timeout	CAPTCHA	5xx
Bright Data	47%	14%	19%	11%	9%
ZenRows	37%	23%	17%	14%	9%
Apify	42%	20%	21%	10%	7%

How to read this page

Success: HTTP 200 with the expected content selector present. Soft blocks count as failures.
p50 / p95: Time to full response, measured from the probe region. Includes provider retries.
Cost / 1k ok: List price per 1,000 requests ÷ sample success rate.
Reliability: 0–100. 60% success, 25% p95 stability, 15% error diversity. Formula →

All results, deltas and error shares are simulated. Region and target selections label the scenario; only JS has a modeled performance adjustment.

How to read these results →