```yaml
product: AlterLab
title: "Web Scraping API Benchmarks 2026 — Real Success Rates, Speed & Cost Data | AlterLab"
comparison_context: "AlterLab is an alternative to Firecrawl, ScrapingBee, and Bright Data."
source_url: https://alterlab.io/benchmark
```

Anti-Bot Difficulty Index: 37,746 verified scrape encounters as of November 1, 2026

# Web Scraping API Benchmarks

Original performance data from AlterLab's 5-tier scraping infrastructure — success rates by website category, geographic latency, success rates by protection type, and cost comparisons. Updated June 2026.

98%+

Overall Success

$0.0002

Lowest Cost

200ms

Fastest Response

6+

Protection Types

## Reproducible Benchmark

Measured on 2026-10-22 across a stated set of 24 public pages (3 attempts each, 72 requests total) using the public API with default tier escalation and caching off. The target set, the runner script and every raw attempt are published, so anyone can re-run it and check the numbers.

91.7%

Success rate

7.7s

p50 latency

44.5s

p95 latency

$1.31

Cost per 1k results

| Category | Targets | Successes | Success rate | p50 | p95 | Cost / 1k |
| --- | --- | --- | --- | --- | --- | --- |
| E-commerce | 4 | 9/12 | 75.0% | 9.2s | 49.0s | $2.27 |
| News & Media | 4 | 12/12 | 100.0% | 10.7s | 45.7s | $1.75 |
| Finance | 4 | 12/12 | 100.0% | 8.2s | 14.9s | $0.23 |
| Government | 4 | 12/12 | 100.0% | 2.5s | 5.8s | $0.20 |
| Real Estate | 4 | 9/12 | 75.0% | 5.9s | 8.3s | $1.57 |
| Jobs & Directories | 4 | 12/12 | 100.0% | 8.6s | 48.3s | $2.17 |

[Raw results (CSV)](https://alterlab.io/benchmark/results.csv)[Raw results (JSON)](https://alterlab.io/benchmark/results.json)[Target set (JSON)](https://alterlab.io/benchmark/targets.json)

Reproduce it with your own API key:

```
cd web && ALTERLAB_API_KEY=... npm run benchmark:public
```

Success means the API returned the page with an upstream 2xx/3xx status, at least 1,000 characters of content and the target's expected text marker. Latency is client wall-clock time for successful requests. Cost per 1k is the billed cost of all attempts divided by successful results. Full definitions are on the [reproducible benchmark methodology](https://alterlab.io/benchmark/methodology#reproducible-benchmark) page.

## Anti-Bot Difficulty Index

A neutral, dated, vendor-level read on how hard each anti-bot system is to bypass right now — and whether it is trending harder or easier. Computed continuously from real production scrape outcomes, not a lab benchmark.

37,746 verified scrape encounters as of November 1, 2026, 90-day window. [Full methodology](https://alterlab.io/benchmark/methodology).

| Vendor | Difficulty Score | Trend | Sample Size |
| --- | --- | --- | --- |
| [_cf_chl](https://alterlab.io/benchmark/vendor/cf-chl) | 100.0 | 0.0 | 74 |
| [challenge-platform](https://alterlab.io/benchmark/vendor/challenge-platform) | 100.0 | 0.0 | 64 |
| [fastly_client_identity](https://alterlab.io/benchmark/vendor/fastly-client-identity) | 100.0 | 0.0 | 59 |
| [imperva](https://alterlab.io/benchmark/vendor/imperva) | 100.0 | 0.0 | 39 |
| [cf-turnstile](https://alterlab.io/benchmark/vendor/cf-turnstile) | 100.0 | 0.0 | 20 |
| [small_block_page](https://alterlab.io/benchmark/vendor/small-block-page) | 99.4 | 0.0 | 330 |
| [empty_response](https://alterlab.io/benchmark/vendor/empty-response) | 99.0 | 0.0 | 959 |
| [imperva_soft](https://alterlab.io/benchmark/vendor/imperva-soft) | 96.8 | 0.0 | 280 |
| [fastly_ip_block](https://alterlab.io/benchmark/vendor/fastly-ip-block) | 95.8 | 0.0 | 264 |
| [datadome](https://alterlab.io/benchmark/vendor/datadome) | 95.3 | 0.0 | 2,582 |
| [access_denied](https://alterlab.io/benchmark/vendor/access-denied) | 81.5 | 0.0 | 1,161 |
| [kasada](https://alterlab.io/benchmark/vendor/kasada) | 71.4 | 0.0 | 28 |
| [akamai_waf](https://alterlab.io/benchmark/vendor/akamai-waf) | 69.4 | 0.0 | 36 |
| [cloudflare_blocked](https://alterlab.io/benchmark/vendor/cloudflare-blocked) | 62.9 | 0.0 | 1,420 |
| [soft_block](https://alterlab.io/benchmark/vendor/soft-block) | 62.4 | 0.0 | 13,188 |
| [cloudflare_js](https://alterlab.io/benchmark/vendor/cloudflare-js) | 61.7 | 0.0 | 1,155 |
| [cloudflare_turnstile](https://alterlab.io/benchmark/vendor/cloudflare-turnstile) | 59.7 | 0.0 | 3,210 |
| [perimeterx](https://alterlab.io/benchmark/vendor/perimeterx) | 55.8 | 0.0 | 120 |
| [captcha](https://alterlab.io/benchmark/vendor/captcha) | 55.4 | 0.0 | 1,440 |
| [blocked](https://alterlab.io/benchmark/vendor/blocked) | 54.4 | 0.0 | 1,734 |
| [radware](https://alterlab.io/benchmark/vendor/radware) | 51.7 | 0.0 | 29 |
| [akamai](https://alterlab.io/benchmark/vendor/akamai) | 49.3 | 0.0 | 136 |
| [imperva_js](https://alterlab.io/benchmark/vendor/imperva-js) | 48.9 | 0.0 | 94 |
| [cloudflare](https://alterlab.io/benchmark/vendor/cloudflare) | 41.3 | 0.0 | 92 |
| [cloudflare_managed_challenge](https://alterlab.io/benchmark/vendor/cloudflare-managed-challenge) | 35.8 | 0.0 | 8,422 |
| [js_render_required](https://alterlab.io/benchmark/vendor/js-render-required) | 34.1 | 0.0 | 507 |
| [shape](https://alterlab.io/benchmark/vendor/shape) | 30.8 | 0.0 | 39 |
| [aws_waf](https://alterlab.io/benchmark/vendor/aws-waf) | 20.1 | 0.0 | 264 |

Vendor-level aggregate only — no per-domain data, no cost figures, no raw bypass-success percentage. Embed this data (CC BY 4.0) via the [embeddable widget](https://alterlab.io/benchmark/embed). Live success, latency and cost by anti-bot class and tier: [reliability feed](https://alterlab.io/.well-known/alterlab-reliability.json).

## Anti-Bot Success Rates by Tier

Each tier adds capability and cost. AlterLab auto-escalates from the cheapest tier, so you only pay for what's needed.

1

Curl

Tier 1

on unprotected sites95%

200–500ms

$0.0002/req

Ultra-fast curl binary-Static HTML, public APIs, RSS feeds

2

HTTP

Tier 2

on basic protection85%

500–2,000ms

$0.0003/req

HTTPX with TLS fingerprinting & HTTP/2-Sites with IP-based blocking, rate limits

3

Stealth

Tier 3

on JS-challenge sites90%

2–5s

$0.002/req

curl_cffi with Chrome browser impersonation-JS challenge-protected sites, basic bot detection

4

Browser

Tier 4

on heavily protected sites92%

5–15s

$0.004/req

Playwright browser automation-Heavily protected sites, SPAs, advanced bot detection

5

Captcha

Tier 5

on hardened targets88%

10–30s

$0.02/req

Browser with AI-powered challenge resolution-Hardened targets, challenge resolution, image/audio challenges

Auto-escalation achieves 98%+ overall success. AlterLab starts at the cheapest tier and automatically escalates through higher tiers only when a lower tier fails. You pay for the tier that succeeds, not the maximum.

## Success Rates by Website Category

Aggregated success rates across 6 major website categories. These are AlterLab's own operational estimates, not third-party numbers — see the live, independently verifiable Anti-Bot Difficulty Index above for the dataset backed by real-time production telemetry.

Methodology: success rate = requests returning valid HTML content / total requests. Data from January–June 2026 production traffic.

E-commerce

Product pages, listings, pricing data

4.2M requests99%

1400ms avg

T1–T3

News & Media

Articles, RSS feeds, live news sites

2.8M requests98%

680ms avg

T1–T2

Financial Data

Market data, filings, public financial records

1.1M requests96%

1900ms avg

T2–T3

Government & Public Records

Agency portals, public databases, regulatory filings

520K requests97%

2100ms avg

T1–T2

Real Estate

Property listings, MLS portals, address databases

890K requests94%

3200ms avg

T3–T4

Job Boards & Directories

Career listings, business directories, review sites

1.5M requests93%

2800ms avg

T2–T4

## Response Latency by Geographic Region

End-to-end latency from API request to response delivery, measured across AlterLab's proxy node network. Proxy nodes are co-located in each region to minimize hops.

Methodology: p50 = median response time, p95 = 95th percentile. Measured at Tier 1 (Curl) baseline. Higher tiers add rendering overhead. Data from Q1–Q2 2026.

North America

US, Canada, Mexico

98%

success rate

420ms

p50 (median)

1,850ms

p95

18 proxy node locations in region

Europe

UK, DE, FR, NL, PL + 20 countries

97%

success rate

610ms

p50 (median)

2,400ms

p95

24 proxy node locations in region

Asia-Pacific

JP, SG, AU, IN, KR, TW

95%

success rate

890ms

p50 (median)

3,700ms

p95

12 proxy node locations in region

Latin America

BR, AR, CL, CO, MX

93%

success rate

1100ms

p50 (median)

4,200ms

p95

8 proxy node locations in region

## Success Rates by Protection Type

How AlterLab performs against each class of website protection mechanism, measured against real production targets. Success rate is the percentage of requests that return valid content when the protection type is present.

Methodology: protection type classified by AlterLab's detection layer before tier selection. Success measured at the first tier that handles the detected protection class. Data from Q1–Q2 2026 production traffic.

Static IP Blocking

Rate limits, IP reputation databases, CIDR blacklists

Tier 298%

Method: Residential proxy rotation with backoff

JavaScript Challenges

Browser environment verification, execution-time JS validation

Tier 394%

Method: Chrome impersonation with full JS execution context

Behavioral Detection

Mouse movement analysis, interaction timing, session fingerprinting

Tier 491%

Method: Full browser automation with human-like event simulation

TLS Fingerprint Matching

JA3/JA4 TLS fingerprint inspection, HTTP/2 stack analysis

Tier 2–396%

Method: Native browser TLS stack with accurate cipher suite ordering

Token-Based Challenges

Single-use access tokens, proof-of-work systems, invisible challenge gates

Tier 4–589%

Method: Browser execution with automated token extraction

Image & Audio Challenges

Grid-selection puzzles, distorted text, audio transcription gates

Tier 587%

Method: AI-powered challenge resolution with multi-modal solver

## Success Rates by Domain

Per-domain performance across 20 frequently scraped sites. Success rate, median latency, cost per successful request, and the recommended tier for each target.

Methodology: success rate = requests returning valid HTML / total requests per domain. Median latency is end-to-end including proxy routing. Cost per success reflects the tier that typically resolves the domain. Data from Q1–Q2 2026 production traffic. Domains with fewer than 50K requests excluded.

| Domain | Category | Success Rate | Median Latency | Cost / Success | Rec. Tier |
| --- | --- | --- | --- | --- | --- |
| amazon.com | E-commerce | 99% | 980ms | $0.0003 | T2 |
| reddit.com | Social / Community | 98% | 620ms | $0.0002 | T1–T2 |
| linkedin.com | Professional Network | 91% | 4,100ms | $0.004 | T4 |
| zillow.com | Real Estate | 95% | 2,800ms | $0.002 | T3 |
| indeed.com | Job Board | 93% | 3,200ms | $0.002 | T3 |
| glassdoor.com | Job Board | 90% | 5,200ms | $0.004 | T4 |
| yelp.com | Reviews / Directories | 94% | 2,400ms | $0.002 | T3 |
| airbnb.com | Travel / Marketplace | 92% | 4,800ms | $0.004 | T4 |
| booking.com | Travel / Marketplace | 96% | 1,900ms | $0.002 | T3 |
| walmart.com | E-commerce | 97% | 1,400ms | $0.0003 | T2 |
| etsy.com | E-commerce | 99% | 850ms | $0.0002 | T1–T2 |
| realtor.com | Real Estate | 96% | 2,100ms | $0.002 | T2–T3 |
| bloomberg.com | Financial News | 88% | 6,400ms | $0.004 | T4 |
| reuters.com | News & Media | 99% | 590ms | $0.0002 | T1 |
| sec.gov | Government / Financial | 98% | 1,100ms | $0.0002 | T1 |
| trulia.com | Real Estate | 95% | 2,600ms | $0.002 | T3 |
| expedia.com | Travel / Marketplace | 97% | 1,700ms | $0.0003 | T2 |
| yellowpages.com | Business Directory | 98% | 780ms | $0.0002 | T1 |
| crunchbase.com | Business Intelligence | 89% | 5,800ms | $0.004 | T4 |
| g2.com | Reviews / Software | 93% | 3,400ms | $0.002 | T3 |

## Response Time Benchmarks

End-to-end response times including proxy routing, rendering, and content extraction. Measured from API request to response delivery.

| Tier | Method | Typical Range | Visualization |
| --- | --- | --- | --- |
| T1: Curl | Ultra-fast curl binary | 200–500ms |  |
| T2: HTTP | HTTPX with TLS fingerprinting & HTTP/2 | 500–2,000ms |  |
| T3: Stealth | curl_cffi with Chrome browser impersonation | 2–5s |  |
| T4: Browser | Playwright browser automation | 5–15s |  |
| T5: Captcha | Browser with AI-powered challenge resolution | 10–30s |  |

## Cost Comparison

Per-request costs across major scraping API providers. AlterLab's pay-per-request model has no monthly minimums or expiring balance.

| Provider | Basic Scrape | JS Rendering | Captcha Solve | Pricing Model |
| --- | --- | --- | --- | --- |
| AlterLabYou are here | $0.0002 | $0.004 | $0.02 | Pay-per-request |
| ScraperAPI | $0.001 | $0.005 | $0.015 | Monthly subscription |
| Bright Data | $0.001 | $0.01 | $0.03 | Monthly + bandwidth |
| Firecrawl | $0.002 | $0.002 | N/A | Monthly subscription |
| ScrapingBee | $0.001 | $0.005 | $0.015 | Monthly subscription |

Competitor pricing based on publicly available pricing pages as of June 2026. Actual costs may vary by plan and volume.

## Anti-Bot System Coverage

AlterLab's multi-tier architecture handles every major anti-bot system in production today.

JS Challenge Protection (lightweight)

Tier 3

JS Challenge Protection (managed)

Tier 4

Behavioral Bot Detection

Tier 4

Advanced Fingerprint Detection

Tier 4

Enterprise Bot Management

Tier 4–5

Token-based Challenge Systems

Tier 5

Image & Audio Challenge Resolution

Tier 5

WAF Rate Limiting

Tier 3–4

Layer 7 DDoS Protection

Tier 4

## How Auto-Escalation Works

AlterLab automatically routes each request through the optimal tier, minimizing cost while maximizing success rate.

1

### Start Cheap

Every request begins at Tier 1 ($0.0002). If the target responds with valid HTML, you're done.

2

### Detect & Escalate

If blocked, AlterLab detects the anti-bot system and jumps to the right tier — no wasted intermediate attempts.

3

### Pay for Success

You're charged only for the tier that succeeds. Domain intelligence learns optimal tiers for future requests.

## Methodology

For the live Anti-Bot Difficulty Index above, see the dedicated, dated [Difficulty Index methodology page](https://alterlab.io/benchmark/methodology) for the exact scoring formula, trend computation, and minimum sample floor. The sections below describe the separate historical benchmark dataset.

All benchmark data is sourced from aggregated production traffic on AlterLab's scraping infrastructure. No synthetic benchmarks or controlled lab environments — these are real request outcomes from paying customers, anonymized and aggregated by category, region, and protection type.

### Success Rate Definition

Success rate = requests returning valid HTML content / total requests attempted at a given tier or against a given protection type. The “overall 98%+ success rate” reflects AlterLab's auto-escalation system: the system routes requests through progressively more capable tiers until a valid response is obtained.

### Website Category Classification

Categories are assigned using AlterLab's domain intelligence layer based on domain metadata and content type signals. Sample sizes reflect the number of requests classified in each category from January–June 2026 production traffic. Categories with fewer than 100K requests are excluded to ensure statistical confidence.

### Geographic Latency

Latency is measured end-to-end from API request receipt to response delivery, including proxy routing, DNS resolution, TLS handshake, page rendering (where applicable), and content extraction. p50 is the median response time; p95 is the 95th percentile. Baseline measurements use Tier 1 (Curl) — higher tiers add rendering overhead. Proxy nodes are deployed in each reported region.

### Protection Type Classification

AlterLab's detection layer classifies the primary protection mechanism encountered during each request. When multiple protection types are present, the most restrictive is used for classification. Success rate for a protection type reflects performance specifically against that class of challenge at the appropriate handling tier.

### Competitor Pricing

Competitor pricing data is sourced from publicly available pricing pages as of June 2026 and may not reflect negotiated enterprise rates or volume discounts.

Last updated: October 2026

## More Web Scraping API Resources

Part of the [Web Scraping API Documentation](https://alterlab.io/docs) cluster

[Web Scraping API DocumentationComplete API reference with 5-tier auto-escalation — Curl to challenge resolution.Pillar page](https://alterlab.io/docs)[Web Scraping API Comparison 2026Side-by-side comparison with ScraperAPI, Firecrawl, Bright Data, and Apify.](https://alterlab.io/vs)[AlterLab for AI AgentsUse web scraping in Claude, GPT-4, and LangChain agent workflows.](https://alterlab.io/for-ai-agents)[Automatic Website Compatibility APIHandles bot-protected websites, JavaScript challenges, and complex protections automatically.](https://alterlab.io/smart-rendering-api)[JavaScript Rendering APIFull browser rendering for SPAs, React, and dynamic content.](https://alterlab.io/javascript-rendering-api)[Search, Crawl, and Map PipelineChain Search → Crawl → Map operations in a single API workflow.](https://alterlab.io/platform)[Pay-as-You-Go PricingFrom $0.0002/request. Balance never expires — no monthly subscription.](https://alterlab.io/pricing)

Live test

## Test your own URL

See how AlterLab performs against any URL you care about — no account required. The playground runs a real scrape and returns timing, tier used, and the extracted content.

[Open playground](https://alterlab.io/playground)

## Your first scrape. Sixty seconds.

Up to 5,000 free requests — about 5,000 Light, 500 Protected or 250 Dynamic requests. No credit card. 
Just a POST request.

terminal

curl -X POST https://api.alterlab.io/v1/scrape \

-H "X-API-Key: YOUR_KEY" \

-H "Content-Type: application/json" \

-d '{"url": "https://example.com", "formats": ["markdown"]}'

[Start building free](https://alterlab.io/signin)

No credit card required · Up to 5,000 free requests (about 5,000 Light, 500 Protected or 250 Dynamic requests) · Balance never expires