Speed scores only matter if you know which data they come from. Verdict first, then the rubric.
Quick answer
Google PageSpeed Insights scores 90/100 and wins for most people: it is free, needs no account, and is the only tool that shows both Lighthouse lab data and CrUX field data — the real-world 75th-percentile metrics Google references. For the deepest single-test diagnosis (filmstrips, waterfalls, multi-location runs), WebPageTest is the runner-up. If you need to watch speed continuously and get alerted on regressions, choose DebugBear.
The ranking
| Rank | Tool | Best for | Entry price | SR Score |
|---|---|---|---|---|
| 1 | Google PageSpeed Insights | Free lab + CrUX field checks | Free | 90 |
| 2 | WebPageTest (Catchpoint) | Deep one-off diagnosis | Free; Pro from ~$15/mo | 88 |
| 3 | DebugBear | Continuous monitoring + alerts | From ~$49/mo | 86 |
| 4 | GTmetrix | Cheap scheduled monitoring | Free; PRO from ~$5–10/mo | 82 |
| 5 | Lighthouse (Chrome DevTools) | Free local lab audits | Free | 80 |
How to choose
Match the tool to the question. “Do my real users have a speed problem?” → PageSpeed Insights or any RUM source, because only field data (CrUX) reflects real users and only field data feeds Google’s page-experience signal. “Why is this specific page slow?” → WebPageTest or Chrome DevTools, where waterfalls and filmstrips expose the exact blocking resource, slow server response, or layout shift. “Did our last deploy make things worse, and will it again?” → DebugBear or GTmetrix, which schedule tests and alert on regressions.
The mistake most teams make is optimizing for a lab score they can control while ignoring the field score Google actually uses. Lab tools are for diagnosis; field data is for deciding whether to act. The strongest workflow uses both: watch CrUX to know if you have a problem, then drop into a lab tool to find and fix the cause, then confirm the fix moved the field number a few weeks later.
Methodology
Weights: Measurement accuracy & evidence 30, Diagnostic depth 25, Value for money 20, Support & docs 15, Transparency 10.
Accuracy leads, and we reward tools that show field data (CrUX/RUM — what Google actually measures) alongside lab data. Diagnostic depth covers waterfalls, filmstrips, opportunity lists, and metric-level detail (LCP/INP/CLS). Value is at the plan a buyer needs.
Google PageSpeed Insights — 90
Free, authoritative, and unique in combining lab and field data. PSI runs Lighthouse 13.0 (updated October 2025) for lab scores and Opportunities, then layers in CrUX field data at the 75th percentile over a rolling 28-day window. For deciding whether real users experience a speed problem, nothing free is better.
Trade-off: it is a snapshot — no scheduling, history, or alerts on the public tool — and field data needs enough CrUX traffic to display.
| Criterion | Score |
|---|---|
| Measurement accuracy & evidence (30) | 28 |
| Diagnostic depth (25) | 21 |
| Value for money (20) | 20 |
| Support & docs (15) | 13 |
| Transparency (10) | 8 |
WebPageTest (Catchpoint) — 88
The deepest lab diagnostic tool. Run from many locations and devices with controllable connection profiles, and get request waterfalls, a visual filmstrip, and Core Web Vitals breakdowns. Free Starter = 300 runs/month; Pro from about $15/mo (annual) / $18.75 monthly for 1,000 runs.
Trade-off: primarily lab data (CrUX appears in places, but it is not a field-first tool), and its power has a learning curve.
| Criterion | Score |
|---|---|
| Measurement accuracy & evidence (30) | 25 |
| Diagnostic depth (25) | 24 |
| Value for money (20) | 18 |
| Support & docs (15) | 13 |
| Transparency (10) | 8 |
DebugBear — 86
The monitoring choice. Scheduled synthetic Lighthouse tests plus real-user data track LCP/INP/CLS over time and alert on regressions, with waterfalls to diagnose causes. From about $49/month, scaling with sites and frequency.
Trade-off: paid, and overkill for a one-time test versus the free tools.
| Criterion | Score |
|---|---|
| Measurement accuracy & evidence (30) | 26 |
| Diagnostic depth (25) | 22 |
| Value for money (20) | 16 |
| Support & docs (15) | 14 |
| Transparency (10) | 8 |
GTmetrix — 82
Affordable scheduled monitoring with a friendly report. PRO tiers start very low — Micro ~$5/mo, Solo ~$10/mo (annual), Starter ~$20/mo — adding monitored slots, hourly checks, premium locations, and mobile testing.
Trade-off: its scoring leans on Lighthouse lab data; for the CrUX field view you still need PSI or RUM.
| Criterion | Score |
|---|---|
| Measurement accuracy & evidence (30) | 21 |
| Diagnostic depth (25) | 21 |
| Value for money (20) | 18 |
| Support & docs (15) | 12 |
| Transparency (10) | 10 |
Lighthouse (Chrome DevTools) — 80
The free lab engine inside Chrome. Run a Performance audit on any page and get scores, Opportunities, and Diagnostics, plus the Performance panel for INP and main-thread profiling.
Trade-off: lab-only and local — no field data, history, or alerts. It tells you why a page is slow on your machine, not whether real users suffer.
| Criterion | Score |
|---|---|
| Measurement accuracy & evidence (30) | 20 |
| Diagnostic depth (25) | 22 |
| Value for money (20) | 20 |
| Support & docs (15) | 12 |
| Transparency (10) | 6 |
Re-weighting the rank
PageSpeed Insights wins because we weight measurement accuracy and field evidence highest, and it is free and authoritative on both. If you weight diagnostic depth at 40 instead, WebPageTest overtakes it — its waterfalls and multi-location runs are deeper than PSI’s. If you weight ongoing monitoring (not a criterion we scored explicitly, but folded into capability) above everything, DebugBear becomes your #1, because spot-check tools cannot alert you. The ranking is honest about being a general-purpose recommendation; your specific job may justify a different top pick, and the per-criterion tables let you redo the math.
Verification
- Google PageSpeed Insights — Lighthouse 13.0 (Oct 2025) and CrUX field methodology (75th percentile / 28-day window) verified on developers.google.com/speed/docs/insights and release notes.
- WebPageTest — free Starter (300 runs) and Pro from ~$15/mo annual verified on catchpoint.com/pricing and TrustRadius.
- DebugBear — from ~$49/mo monitoring features verified on debugbear.com/pricing.
- GTmetrix — PRO tiers verified on gtmetrix.com/pricing.
- Lighthouse / Chrome DevTools — free, in-browser; documented at developer.chrome.com and Google Search Central.
Related rankings
- Best Core Web Vitals Tools (2026): Scored & Ranked
- Best Redirect Checker Tools (2026): .htaccess & 301 Ranked
- Best Image Optimization Tools (2026): Scored & Ranked
- Best Indexing Tools for SEO (2026): Scored & Ranked
Frequently asked questions
- What is the best free page speed testing tool?
- Google PageSpeed Insights. It is free, requires no account, and uniquely shows both Lighthouse lab data and CrUX field data — the real-world metrics Google uses for the page-experience signal.
- PageSpeed Insights vs. Lighthouse vs. WebPageTest?
- PSI runs Lighthouse plus CrUX field data in one report. Lighthouse alone (in Chrome DevTools) is lab-only and local. WebPageTest gives the deepest lab diagnostics: filmstrips, waterfalls, multi-location and multi-device runs.
- Which tool should I use to monitor speed over time?
- DebugBear or GTmetrix. PSI and WebPageTest are spot-check tools without scheduling and alerting on their lower tiers; DebugBear is built for continuous monitoring from about $49/month, GTmetrix offers cheap monitored slots from a few dollars a month.
- Why do tools report different speed scores?
- They use different test environments, connection throttling, locations, and devices, and lab data differs from field data. Trust CrUX field data for whether real users have a problem; use lab tools to find why.