Benchmarking CAPTCHA Solve Times Across Types with CaptchaAI

Solve time matters. A 10-second difference per solve compounds into hours of delay across thousands of tasks. This guide provides reproducible benchmarks for every CAPTCHA type CaptchaAI supports, plus scripts to run your own measurements.

Typical Solve Times

These benchmarks reflect median times from 100 consecutive solves per type, measured from API submission to result retrieval:

Times vary based on proxy quality, server load, and CAPTCHA complexity.

Factors Affecting Solve Time

Network Latency

Your distance from CaptchaAI's API endpoints adds round-trip time. A 200ms RTT adds ~1 second total (submit + 3–4 polls × 200ms).

Proxy Quality

When using proxies, residential proxies typically produce faster solves than datacenter proxies because they trigger fewer secondary challenges.

CAPTCHA Complexity

reCAPTCHA v2 image challenges vary in difficulty. A "select traffic lights" challenge with fading images takes longer than a simple checkbox verification.

Polling Strategy

Aggressive polling (every 2 seconds) detects results faster but increases API calls. Conservative polling (every 10 seconds) reduces load but adds latency.

Python Benchmark Script

# benchmark_captchaai.py
import os
import time
import statistics
import requests

API_KEY = os.environ.get("CAPTCHAAI_KEY", "YOUR_API_KEY")

def benchmark_solve(method, params, runs=10):
    """Benchmark a CAPTCHA type over multiple runs."""
    times = []
    successes = 0

    for i in range(runs):
        start = time.time()

        # Submit
        submit_params = {"key": API_KEY, "json": "1", **params}
        resp = requests.get("https://ocr.captchaai.com/in.php",
                            params=submit_params)
        result = resp.json()

        if result.get("status") != 1:
            print(f"  Run {i+1}: Submit failed - {result.get('request')}")
            continue

        task_id = result["request"]

        # Poll
        time.sleep(10)
        solved = False
        for _ in range(30):
            poll = requests.get("https://ocr.captchaai.com/res.php", params={
                "key": API_KEY, "action": "get",
                "id": task_id, "json": "1",
            })
            poll_result = poll.json()

            if poll_result.get("status") == 1:
                elapsed = time.time() - start
                times.append(elapsed)
                successes += 1
                solved = True
                print(f"  Run {i+1}: {elapsed:.1f}s")
                break
            if poll_result.get("request") != "CAPCHA_NOT_READY":
                print(f"  Run {i+1}: Error - {poll_result.get('request')}")
                break

            time.sleep(5)

        if not solved and poll_result.get("request") == "CAPCHA_NOT_READY":
            print(f"  Run {i+1}: Timeout")

    if times:
        return {
            "runs": runs,
            "successes": successes,
            "success_rate": f"{successes/runs*100:.0f}%",
            "median": f"{statistics.median(times):.1f}s",
            "p90": f"{sorted(times)[int(len(times)*0.9)]:.1f}s" if len(times) >= 10 else "N/A",
            "min": f"{min(times):.1f}s",
            "max": f"{max(times):.1f}s",
        }
    return {"error": "No successful solves"}

# Benchmark reCAPTCHA v2
print("=== reCAPTCHA v2 ===")
recaptcha_v2 = benchmark_solve("userrecaptcha", {
    "method": "userrecaptcha",
    "googlekey": "6Le-wvkSAAAAAPBMRTvw0Q4Muexq9bi0DJwx_mJ-",
    "pageurl": "https://www.google.com/recaptcha/api2/demo",
})
print(recaptcha_v2)

# Benchmark Image/OCR
print("\n=== Image CAPTCHA ===")
# Submit a base64-encoded test image
import base64
# Use a test image for benchmarking
image_result = benchmark_solve("post", {
    "method": "base64",
    "body": "BASE64_ENCODED_CAPTCHA_IMAGE",
})
print(image_result)

JavaScript Benchmark Script

// benchmark_captchaai.js
const axios = require('axios');

const API_KEY = process.env.CAPTCHAAI_KEY || 'YOUR_API_KEY';

async function benchmarkSolve(params, runs = 10) {
  const times = [];
  let successes = 0;

  for (let i = 0; i < runs; i++) {
    const start = Date.now();

    // Submit
    const submit = await axios.get('https://ocr.captchaai.com/in.php', {
      params: { key: API_KEY, json: '1', ...params },
    });

    if (submit.data.status !== 1) {
      console.log(`  Run ${i + 1}: Submit failed - ${submit.data.request}`);
      continue;
    }

    // Poll
    await new Promise(r => setTimeout(r, 10000));
    let solved = false;

    for (let j = 0; j < 30; j++) {
      const poll = await axios.get('https://ocr.captchaai.com/res.php', {
        params: { key: API_KEY, action: 'get', id: submit.data.request, json: '1' },
      });

      if (poll.data.status === 1) {
        const elapsed = (Date.now() - start) / 1000;
        times.push(elapsed);
        successes++;
        solved = true;
        console.log(`  Run ${i + 1}: ${elapsed.toFixed(1)}s`);
        break;
      }
      if (poll.data.request !== 'CAPCHA_NOT_READY') {
        console.log(`  Run ${i + 1}: Error - ${poll.data.request}`);
        break;
      }
      await new Promise(r => setTimeout(r, 5000));
    }

    if (!solved) console.log(`  Run ${i + 1}: Timeout`);
  }

  if (times.length === 0) return { error: 'No successful solves' };

  times.sort((a, b) => a - b);
  return {
    runs, successes,
    successRate: `${((successes / runs) * 100).toFixed(0)}%`,
    median: `${times[Math.floor(times.length / 2)].toFixed(1)}s`,
    min: `${times[0].toFixed(1)}s`,
    max: `${times[times.length - 1].toFixed(1)}s`,
  };
}

(async () => {
  console.log('=== reCAPTCHA v2 ===');
  const v2 = await benchmarkSolve({
    method: 'userrecaptcha',
    googlekey: '6Le-wvkSAAAAAPBMRTvw0Q4Muexq9bi0DJwx_mJ-',
    pageurl: 'https://www.google.com/recaptcha/api2/demo',
  });
  console.log(v2);

  console.log('\n=== Cloudflare Turnstile ===');
  const turnstile = await benchmarkSolve({
    method: 'turnstile',
    sitekey: 'YOUR_TURNSTILE_SITEKEY',
    pageurl: 'https://example.com',
  });
  console.log(turnstile);
})();

Interpreting Results

Median vs P90

• Median — The typical experience. Use this for capacity planning.
• P90 — The "worst case" for 9 out of 10 solves. Use this for timeout settings.
• P99 — The true edge case. Set your maximum timeout higher than this.

Optimal Polling Intervals

Based on the benchmarks above, configure polling per CAPTCHA type:

Troubleshooting

FAQ

Do solve times vary by time of day?

Yes. Peak hours (US business hours) may show slightly longer solve times due to higher demand. Run benchmarks at different times for a complete picture.

Should I use this data for production timeout settings?

Use P90 values for standard timeouts and P99 for maximum timeouts. Add a 20% buffer to account for network variability.

How do proxy settings affect solve times?

If you supply a proxy, CaptchaAI routes the solve through it. Slow or distant proxies can add 5–15 seconds to total solve time. For benchmarking, test with and without proxies separately.

Related Articles

• Solve Multiple Captcha Types Workflow
• Captchaai Supported Captcha Types Feature Matrix

Next Steps

Run the benchmark scripts against your own account and CAPTCHA targets — get your CaptchaAI API key.

Related guides:

• Optimizing Speed and Cost
• CaptchaAI Latency Optimization
• Parallel CAPTCHA Solving