Next.js

Dev Time

Setup

Dev Time benchmarks use each meta-framework’s starter-* package.

Each starter follows its framework’s CLI or documented setup, with the selected options and adjustments recorded in the methodology.

See Default Benchmark Setup for the shared benchmark environment.

Dependency Counts

Methodology: Dependency Counts.

Measuring a meta-framework’s dependencies isn’t as straightforward as you might think. You could:

• Measure just the meta-framework package itself
• Maybe a full project setup

Both approaches have flaws. Some meta-frameworks need many other packages to work. A full project setup can include more packages than necessary.

To provide a clearer picture, we run the dependency-count tests three times, starting small and expanding in scope. In this order, we:

1. Start with the meta-framework package itself, for example, @sveltejs/kit
2. We then expand this to the meta-framework and its first-party packages. Looking at @sveltejs/kit again, this means any package maintained by the meta-framework authors that’s needed to run a project, for example, the svelte package.
3. Finally, we count the dependencies for the entire starter-* project. Their CLI or documentation recommends this setup by default, so this is the default most people will start with.

Core Package · Production Dependencies

Latest · v16.4.36No change from previous

Core Package · Production Dependencies data

Framework version	Dependencies
16.1.1	6
16.2.12	6
16.3.8	6
16.4.3	6

The latest release is shown as a filled point; earlier releases use outlined points.

Core Package · Development Dependencies

Latest · v16.4.30No change from previous

Core Package · Development Dependencies data

Framework version	Dependencies
16.1.1	223
16.2.12	0
16.3.8	0
16.4.3	0

The latest release is shown as a filled point; earlier releases use outlined points.

Core Package · All Dependencies

Latest · v16.4.36No change from previous

Core Package · All Dependencies data

Framework version	Dependencies
16.1.1	229
16.2.12	6
16.3.8	6
16.4.3	6

The latest release is shown as a filled point; earlier releases use outlined points.

Starter Project · Production Dependencies

Latest · v16.4.33No change from previous

Starter Project · Production Dependencies data

Framework version	Dependencies
16.1.1	3
16.2.12	3
16.3.8	3
16.4.3	3

The latest release is shown as a filled point; earlier releases use outlined points.

Starter Project · Development Dependencies

Latest · v16.4.36No change from previous

Starter Project · Development Dependencies data

Framework version	Dependencies
16.1.1	6
16.2.12	6
16.3.8	6
16.4.3	6

The latest release is shown as a filled point; earlier releases use outlined points.

Starter Project · All Dependencies

Latest · v16.4.3424−2 from previous

Starter Project · All Dependencies data

Framework version	Dependencies
16.1.1	421
16.2.12	423
16.3.8	426
16.4.3	424

The latest release is shown as a filled point; earlier releases use outlined points.

Starter Project · Duplicate Dependencies

Latest · v16.4.312−2 from previous

Starter Project · Duplicate Dependencies data

Framework version	Duplicates
16.1.1	12
16.2.12	14
16.3.8	14
16.4.3	12

The latest release is shown as a filled point; earlier releases use outlined points.

Node Modules Size

Methodology: Node Modules Size.

Node Modules Size

Latest · v16.4.3463.42 MB+10.70 MB from previous

Node Modules Size data

Framework version	MB
16.1.1	560.05 MB
16.2.12	451.10 MB
16.3.8	452.72 MB
16.4.3	463.42 MB

The latest release is shown as a filled point; earlier releases use outlined points.

Dependency Install Size

Latest · v16.4.3405.08 MB+8.82 MB from previous

Dependency Install Size data

Framework version	MB
16.1.1	506.22 MB
16.2.12	395.87 MB
16.3.8	396.25 MB
16.4.3	405.08 MB

The latest release is shown as a filled point; earlier releases use outlined points.

Build and Install Times

Methodology: Build and Install Times.

Median Install Time

Latest · v16.4.36.11s−0.31s from previous

Median Install Time data

Framework version	Seconds
16.1.1	6.75s
16.2.12	5.85s
16.3.8	6.42s
16.4.3	6.11s

The latest release is shown as a filled point; earlier releases use outlined points.

Median Cold Build Time

Latest · v16.4.35.71s−1.08s from previous

Median Cold Build Time data

Framework version	Seconds
16.1.1	5.02s
16.2.12	5.19s
16.3.8	6.79s
16.4.3	5.71s

The latest release is shown as a filled point; earlier releases use outlined points.

Median Warm Build Time

Latest · v16.4.32.35s−0.41s from previous

Median Warm Build Time data

Framework version	Seconds
16.1.1	5.04s
16.2.12	5.01s
16.3.8	2.76s
16.4.3	2.35s

The latest release is shown as a filled point; earlier releases use outlined points.

Dev Server Startup

Methodology: Dev Server Startup.

Median Dev Server Startup Time

Latest · v16.4.32.05s−0.13s from previous

Median Dev Server Startup Time data

Framework version	Seconds
16.1.1	2.64s
16.2.12	2.18s
16.3.8	2.18s
16.4.3	2.05s

The latest release is shown as a filled point; earlier releases use outlined points.

Core-JS Polyfills

Methodology: Core-JS Polyfills.

Size of JS Files Containing Core-JS

Combined size of matching build output files, including any other code in those files.

Latest · v16.4.3109.96 KBNo change from previous

Size of JS Files Containing Core-JS data

Framework version	KB
16.1.1	109.96 KB
16.2.12	109.96 KB
16.3.8	109.96 KB
16.4.3	109.96 KB

The latest release is shown as a filled point; earlier releases use outlined points.

Browser Baseline

Methodology: Browser Baseline.

Version	Baseline	Feature	Year	Features
16.1.1	High	—	2022	46
16.2.12	High	—	2022	47
16.3.8	High	—	2022	47
16.4.3	High	—	2022	47

Run Time

Setup

Run Time benchmarks use each meta-framework’s app-* package.

Currently, each test uses a different route in its app, configured to meet each test’s requirements.

See Default Benchmark Setup for the base set-up. Each test’s specific setup is linked in the methodology with links in each section.

Initial Load

Compare the resources transferred on a first visit to the same small, styled landing page. Sizes include response headers; lower is better. Inline CSS and JavaScript count toward HTML. Other includes images, media, fonts, and remaining resources. See Methodology for collection details.

Initial Load: All Resources by Version

Not enough data yet. This chart needs at least two versions.

Initial Load: Excluding JavaScript by Version

Not enough data yet. This chart needs at least two versions.

Client Side Rendered Tests (Lighthouse)

These CSR browser performance tests use Lighthouse to measure paint timings and a controlled navigation interaction.

Methodology: Client Side Rendered Tests (Lighthouse).

Some frameworks have been moved off their golden path to make client side tests fair when comparing. See the methodology link above for the full details.

Client First Paint (Lighthouse)

Latest · v16.4.3342.00 ms+2.00 ms from previous

Client First Paint (Lighthouse) data

Framework version	Milliseconds
16.1.1	339.00 ms
16.2.12	341.00 ms
16.3.8	340.00 ms
16.4.3	342.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Client FCP (Lighthouse)

Latest · v16.4.3342.21 ms+2.31 ms from previous

Client FCP (Lighthouse) data

Framework version	Milliseconds
16.1.1	339.11 ms
16.2.12	341.32 ms
16.3.8	339.90 ms
16.4.3	342.21 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Client Interaction Breakdown by Version (Lighthouse)

Total latency is split into the three consecutive phases reported by Lighthouse.

• The input delay, which starts when the user initiates an interaction with the page, and ends when the event callbacks for the interaction begin to run.
• The processing duration, which consists of the time it takes for event callbacks to run to completion.
• The presentation delay, which is the time it takes for the browser to present the next frame which contains the visual result of the interaction.

• Input delay
• Processing duration
• Presentation delay

Client Interaction Breakdown by Version (Lighthouse) data

Framework version	Total	Input delay	Processing duration	Presentation delay
v16.1.1	17.88 ms	1.49 ms	5.55 ms	10.83 ms
v16.2.12	20.62 ms	1.63 ms	4.35 ms	14.64 ms
v16.3.8	16.80 ms	1.24 ms	5.87 ms	9.69 ms
v16.4.3	19.90 ms	1.63 ms	4.86 ms	13.41 ms

Server Side Rendered Tests (Lighthouse)

These SSR browser performance tests use Lighthouse to measure paint timings and a controlled navigation interaction.

Methodology: Server Side Rendered Tests (Lighthouse).

Server First Paint (Lighthouse)

Latest · v16.4.3373.00 ms+248.00 ms from previous

Server First Paint (Lighthouse) data

Framework version	Milliseconds
16.1.1	134.00 ms
16.2.12	122.00 ms
16.3.8	125.00 ms
16.4.3	373.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Server FCP (Lighthouse)

Latest · v16.4.3372.88 ms+247.80 ms from previous

Server FCP (Lighthouse) data

Framework version	Milliseconds
16.1.1	134.21 ms
16.2.12	121.75 ms
16.3.8	125.08 ms
16.4.3	372.88 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Server Interaction Breakdown by Version (Lighthouse)

Total latency is split into the three consecutive phases reported by Lighthouse.

• The input delay, which starts when the user initiates an interaction with the page, and ends when the event callbacks for the interaction begin to run.
• The processing duration, which consists of the time it takes for event callbacks to run to completion.
• The presentation delay, which is the time it takes for the browser to present the next frame which contains the visual result of the interaction.

• Input delay
• Processing duration
• Presentation delay

Server Interaction Breakdown by Version (Lighthouse) data

Framework version	Total	Input delay	Processing duration	Presentation delay
v16.1.1	20.25 ms	1.49 ms	6.78 ms	11.99 ms
v16.2.12	17.12 ms	1.34 ms	5.77 ms	10.00 ms
v16.3.8	17.52 ms	1.39 ms	7.00 ms	9.13 ms
v16.4.3	14.49 ms	1.36 ms	4.35 ms	8.78 ms

Server Side Throughput Tests

Methodology: Server Side Throughput Tests.

Measures how many requests each framework’s server handler can process per second when rendering a 1,000-row table, and how long each response takes. Requests bypass network overhead to focus on server rendering performance.

Ops/sec

Latest · v16.4.3134

Ops/sec data

Framework version	Operations per second
16.1.1	—
16.2.12	—
16.3.8	—
16.4.3	134

The latest release is shown as a filled point; earlier releases use outlined points.

Body Size

Latest · v16.4.3204.56 KB

Body Size data

Framework version	Kilobytes
16.1.1	—
16.2.12	—
16.3.8	—
16.4.3	204.56 KB

The latest release is shown as a filled point; earlier releases use outlined points.

Median Latency

Median of measured run average latencies; not a pooled request-level median.

Latest · v16.4.37.70 ms

Median Latency data

Framework version	Milliseconds
16.1.1	—
16.2.12	—
16.3.8	—
16.4.3	7.70 ms

The latest release is shown as a filled point; earlier releases use outlined points.

SSR Load Test

Methodology: SSR Load Test.

Measures server throughput and response latency as concurrent connections increase, with each request rendering a 1,000-row table.

This version uses ordinary <a> links to measure server rendering without the extra work of rendering framework router link components.

Compare it with the router link load test to see how framework router link components affect server rendering performance.

Ordinary anchors at /server-side-rendered-plain-links. Historical standard-test measurements are unavailable; new CI runs will populate this series.

Total Measured Requests

Latest · v16.4.34801−1085 from previous

Total Measured Requests data

Framework version	Requests
16.1.1	3798
16.2.12	5877
16.3.8	5886
16.4.3	4801

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 25 Connections

Latest · v16.4.31219.00 ms+80.00 ms from previous

Median Sweep P99 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	1827.00 ms
16.2.12	1175.00 ms
16.3.8	1139.00 ms
16.4.3	1219.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 50 Connections

Latest · v16.4.34269.00 ms+326.00 ms from previous

Median Sweep P99 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	4723.00 ms
16.2.12	4312.00 ms
16.3.8	3943.00 ms
16.4.3	4269.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 100 Connections

Latest · v16.4.34656.00 ms+54.00 ms from previous

Median Sweep P99 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	4744.00 ms
16.2.12	4636.00 ms
16.3.8	4602.00 ms
16.4.3	4656.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 25 Connections

Latest · v16.4.3521.00 ms+107.00 ms from previous

Median Sweep P90 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	652.00 ms
16.2.12	483.00 ms
16.3.8	414.00 ms
16.4.3	521.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 50 Connections

Latest · v16.4.31768.00 ms+572.00 ms from previous

Median Sweep P90 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	2313.00 ms
16.2.12	1488.00 ms
16.3.8	1196.00 ms
16.4.3	1768.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 100 Connections

Latest · v16.4.32283.00 ms+405.00 ms from previous

Median Sweep P90 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	2584.00 ms
16.2.12	1889.00 ms
16.3.8	1878.00 ms
16.4.3	2283.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 25 Connections

Latest · v16.4.3503.00 ms+99.00 ms from previous

Median Sweep P75 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	640.00 ms
16.2.12	468.00 ms
16.3.8	404.00 ms
16.4.3	503.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 50 Connections

Latest · v16.4.3979.00 ms+199.00 ms from previous

Median Sweep P75 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	1062.00 ms
16.2.12	875.00 ms
16.3.8	780.00 ms
16.4.3	979.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 100 Connections

Latest · v16.4.31014.00 ms+196.00 ms from previous

Median Sweep P75 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	1047.00 ms
16.2.12	902.00 ms
16.3.8	818.00 ms
16.4.3	1014.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

SSR Router Link Load Test

Methodology: SSR Router Link Load Test.

Measures server throughput and response latency as concurrent connections increase, with each request rendering a 1,000-row table using framework router link components.

Compare it with the plain link load test to see how these components affect server rendering performance. These tests make HTTP requests only; they do not measure browser navigation or prefetching.

Frameworks without a router link component are excluded.

Framework router links at /server-side-rendered. Historical HTTP load results are preserved in this series.

Total Measured Requests

Latest · v16.4.34278−1005 from previous

Total Measured Requests data

Framework version	Requests
16.1.1	3438
16.2.12	5248
16.3.8	5283
16.4.3	4278

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 25 Connections

Latest · v16.4.31400.00 ms+130.00 ms from previous

Median Sweep P99 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	1949.00 ms
16.2.12	1273.00 ms
16.3.8	1270.00 ms
16.4.3	1400.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 50 Connections

Latest · v16.4.34758.00 ms+414.00 ms from previous

Median Sweep P99 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	4745.00 ms
16.2.12	4612.00 ms
16.3.8	4344.00 ms
16.4.3	4758.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P99 Latency at 100 Connections

Latest · v16.4.34742.00 ms+149.00 ms from previous

Median Sweep P99 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	4746.00 ms
16.2.12	4529.00 ms
16.3.8	4593.00 ms
16.4.3	4742.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 25 Connections

Latest · v16.4.3569.00 ms+100.00 ms from previous

Median Sweep P90 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	714.00 ms
16.2.12	513.00 ms
16.3.8	469.00 ms
16.4.3	569.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 50 Connections

Latest · v16.4.32234.00 ms+650.00 ms from previous

Median Sweep P90 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	2412.00 ms
16.2.12	1856.00 ms
16.3.8	1584.00 ms
16.4.3	2234.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P90 Latency at 100 Connections

Latest · v16.4.32445.00 ms+470.00 ms from previous

Median Sweep P90 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	2775.00 ms
16.2.12	2127.00 ms
16.3.8	1975.00 ms
16.4.3	2445.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 25 Connections

Latest · v16.4.3549.00 ms+94.00 ms from previous

Median Sweep P75 Latency at 25 Connections data

Framework version	Milliseconds
16.1.1	700.00 ms
16.2.12	504.00 ms
16.3.8	455.00 ms
16.4.3	549.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 50 Connections

Latest · v16.4.31107.00 ms+242.00 ms from previous

Median Sweep P75 Latency at 50 Connections data

Framework version	Milliseconds
16.1.1	1109.00 ms
16.2.12	934.00 ms
16.3.8	865.00 ms
16.4.3	1107.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.

Median Sweep P75 Latency at 100 Connections

Latest · v16.4.31065.00 ms+200.00 ms from previous

Median Sweep P75 Latency at 100 Connections data

Framework version	Milliseconds
16.1.1	1088.00 ms
16.2.12	925.00 ms
16.3.8	865.00 ms
16.4.3	1065.00 ms

The latest release is shown as a filled point; earlier releases use outlined points.