Choose a version:
35% The original file has 35099 bytes (34.3k) and is available from the project website.
There you can find the official minified version, too, which brings down the size to 12122 bytes (11.8k, 35%).

After GZIP compression these minified files vary in size:
Boot
  4577 bytes (4.5k)
CDN
Baidu
  4116 bytes (4.0k)
CDN
gzip -6 (default)
  4110 bytes (4.0k)
local copy
gzip -9
  4103 bytes (4.0k)
local copy
cdnjs
  4098 bytes (4.0k)
CDN
unpkg
  4090 bytes (4.0k)
CDN
libdeflate -12
  3999 bytes (3.9k)
local copy
zultra
  3996 bytes (3.9k)
local copy
7zip -mx=9 -tgzip
  3994 bytes (3.9k)
local copy
kzip -s0 -rn -b0
  3992 bytes (3.9k)
local copy
pigz -11 -n
  3985 bytes (3.9k)
local copy
Zopfli
  3976 bytes (3.9k)
local copy

perma-link to the smallest file on my server:
http://minime.stephan-brumme.com/files/underscorejs/underscore-1.3.0.min.js (or via HTTPS)

You will automatically get the smallest UnderscoreJS 1.3.0 file, ETag caching is available and
if your browser doesn't support GZIP decompression then the uncompressed version will be sent.

Currently best Zopfli settings

Save 114 bytes by using my UnderscoreJS 1.3.0 Zopfli version instead of the best available CDN (2.87% smaller than unpkg, 3976 vs. 4090 bytes):
You can use my super-compressed files for whatever purpose you like as long as you respect the library's original license agreement.
There are no restrictions from my side - but please avoid hot-linking if you run a high-traffic website.

These command-line settings yielded the best compression ratio so far (Linux version of zopfli-krzymod):
zopfli --i1000000 --mb8 --mls32 --bsr8 --lazy --ohh

(found February 26, 2020)
Description Value Parameter
iterations 1000000  --i1000000
maximum blocks 8  --mb8
maximum length score 32  --mls32
block splitting recursion 8  --bsr8
lazy matching in LZ77 yes  --lazy
optimized Huffman headers yes  --ohh
initial random W for iterations 1  --rw1
initial random Z for iterations 2  --rz2

Verify file integrity

After decompression, my uncompressed files are identical to the original ones:

MD5:
curl --silent --compressed https://raw.githubusercontent.com/jashkenas/underscore/1.3.0/underscore-min.js --location | md5sum
479cbbe3b04152bf4b6c7531f6bf200d  -
curl --silent --compressed https://minime.stephan-brumme.com/files/underscorejs/underscore-1.3.0.min.zopfli.js.gz | md5sum
479cbbe3b04152bf4b6c7531f6bf200d  -

SHA1:
curl --silent --compressed https://raw.githubusercontent.com/jashkenas/underscore/1.3.0/underscore-min.js --location | sha1sum
86752ba2e4f3855a8a5e45d5f07567ea70a9cc98  -
curl --silent --compressed https://minime.stephan-brumme.com/files/underscorejs/underscore-1.3.0.min.zopfli.js.gz | sha1sum
86752ba2e4f3855a8a5e45d5f07567ea70a9cc98  -

These CDNs send you the original file:
CDN Size (compressed) MD5 (uncompressed) Timestamp
Boot 4577 bytes 479cbbe3b04152bf4b6c7531f6bf200d (invalid)
cdnjs 4098 bytes 479cbbe3b04152bf4b6c7531f6bf200d (invalid)
unpkg 4090 bytes 479cbbe3b04152bf4b6c7531f6bf200d July 11, 2016 @ 16:49

And some CDNs send you a different file:
CDN Size (compressed) MD5 (uncompressed) Comment / Diff Timestamp
Baidu 4116 bytes 657ef8f863a48206d0ce87e73c7ca333 only whitespaces differ (invalid)

Note: only the MD5 hashes are shown to keep things simple.

Other Versions

Available UnderscoreJS versions at minime.stephan-brumme.com:

1.13.6, 1.13.5, 1.13.4, 1.13.3, 1.13.2, 1.13.1, 1.13.0,
1.12.1, 1.12.0,
1.11.0,
1.10.2, 1.10.1, 1.10.0,
1.9.2, 1.9.1, 1.9.0,
1.8.3, 1.8.2, 1.8.1,
1.7.0,
1.6.0,
1.5.2, 1.5.0,
1.4.4, 1.4.3, 1.4.2, 1.4.1, 1.4.0,
1.3.3, 1.3.2, 1.3.1, 1.3.0,
1.2.4, 1.2.3, 1.2.2, 1.2.1, 1.2.0,
1.1.7, 1.1.5, 1.1.4, 1.1.3, 1.1.2, 1.1.1, 1.1.0,
1.0.4, 1.0.3, 1.0.2, 1.0.1, 1.0.0,
0.6.0,
0.5.7, 0.5.6, 0.5.5, 0.5.4, 0.5.3, 0.5.2, 0.5.1, 0.5.0,
0.4.7, 0.4.6, 0.4.5, 0.4.4, 0.4.3, 0.4.2, 0.4.1, 0.4.0,
0.3.3, 0.3.2, 0.3.1, 0.3.0,
0.2.0,
0.1.1, 0.1.0

The project site contains an overview how well these versions were compressed.
Other interesting projects are AngularJS, BackboneJS, Bootstrap, D3, Dojo, Ember, jQuery, Knockout, lodash, React, Socket.IO, ThreeJS and Vue.

Changelog

Best Zopfli parameters so far:
Size Improvement Parameters Found
3976 bytes -1 byte zopfli --i1000000 --mls32 --bsr8 --lazy --ohh February 26, 2020 @ 15:19
3977 bytes -1 byte zopfli --i10000 --mls32 --bsr8 --lazy --ohh November 20, 2015 @ 16:56
3978 bytes -1 byte zopfli --i1000 --mls8192 --bsr2 --lazy --ohh September 8, 2015 @ 10:57
3979 bytes -1 byte zopfli --i1000 --mls512 --bsr8 --lazy --ohh September 8, 2015 @ 10:56
3980 bytes zopfli --i100 --mls4096 --bsr16 --lazy --ohh September 8, 2015 @ 10:32

If there are multiple parameter sets yielding the same compressed size, only the first one found is shown.

Most recent activity on July 20, 2020 @ 12:55.

Heatmaps

This Zopfli heatmap visualizes how compression changes when modifying the --bsr and --mls parameter.
Cell's contents is the best filesize achieved (in bytes, hover with mouse over cells to see number of iterations).

Good parameters are green, bad are red. The best and worst are bold as well.
The brightness of the blue background color indicates how many iterations were processed:
10,000, 100,000 or 1,000,000.
bsr \ mls
2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768
bsr \ mls
2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768
3982 3982 3984 3982 3982 3982 3984 3983 3982 3982 3982 3979 3978 3978 3978
3978 3978 3978 3978 3979 3982 3980 3979 3989 3983 3982 3978 3981 3978 3978
3984 3984 3978 3978 3980 3980 3979 3979 3978 3980 3980 3978 3979 3978 3978
3979 3979 3979 3979 3978 3978 3984 3979 3978 3980 3982 3980 3978 3978 3978
3979 3979 3978 3978 3976 3982 3981 3978 3978 3980 3982 3978 3978 3978 3978
3978 3978 3978 3978 3985 3979 3979 3980 3978 3980 3980 3980 3979 3978 3978
3978 3978 3991 3978 3982 3981 3981 3978 3979 3980 3982 3981 3978 3978 3978
3978 3978 3978 3978 3978 3981 3979 3978 3978 3981 3980 3981 3981 3979 3979
3979 3979 3979 3978 3979 3979 3978 3978 3978 3981 3980 3978 3978 3978 3978
3979 3979 3979 3978 3979 3980 3979 3978 3978 3980 3980 3982 3981 3979 3979
3979 3979 3979 3992 3978 3982 3982 3979 3989 3980 3982 3981 3978 3979 3979
3978 3978 3978 3979 3977 3982 3984 3979 3978 3982 3982 3979 3978 3978 3978
3978 3978 3978 3978 3980 3982 3981 3978 3978 3980 3980 3978 3980 3979 3979
3979 3979 3984 3978 3985 3979 3984 3978 3979 3980 3980 3981 3978 3978 3978
3979 3979 3978 3978 3978 3981 3979 3978 3978 3980 3982 3980 3978 3980 3980
3978 3978 3978 3978 3978 3978 3982 3978 3978 3980 3980 3980 3981 3980 3980
3978 3978 3978 3985 3984 3984 3979 3978 3978 3980 3980 3978 3979 3980 3980
3978 3978 3978 3978 3978 3978 3982 3978 3978 3980 3982 3980 3979 3980 3980
3978 3978 3978 3978 3978 3978 3982 3978 3978 3980 3980 3980 3981 3980 3980
3978 3978 3978 3978 3978 3979 3982 3978 3978 3980 3982 3980 3978 3978 3978
3984 3984 3985 3985 3979 3979 3982 3978 3978 3980 3981 3978 3978 3978 3978
3979 3979 3978 3979 3979 3985 3981 3979 3978 3981 3982 3978 3982 3978 3978
3979 3979 3979 3978 3979 3978 3982 3979 3979 3980 3980 3980 3981 3978 3978

Due to the Monte Carlo design of my search algorithm, not all parameters have reached the same number of iterations yet:
Iterations Min. Bytes Reduction Coverage
100 3979 bytes 100%
1,000 3978 bytes -1 byte 100%
10,000 3977 bytes -1 byte 100%
100,000 3977 bytes 0.58%
1,000,000 3976 bytes -1 byte 0.29%
10,000,000

KZIP has far less options available for tuning/optimization. I only played around with the number of blocks (parameter -n):
Blocks Min. Bytes Compared To Best Zopfli Compared To Best KZIP
3992 bytes +16 bytes (+0.40%)
3993 bytes +17 bytes (+0.43%) +1 byte
4010 bytes +34 bytes (+0.86%) +18 bytes
4007 bytes +31 bytes (+0.78%) +15 bytes
4041 bytes +65 bytes (+1.63%) +49 bytes
4066 bytes +90 bytes (+2.26%) +74 bytes
4097 bytes +121 bytes (+3.04%) +105 bytes
4124 bytes +148 bytes (+3.72%) +132 bytes
4158 bytes +182 bytes (+4.58%) +166 bytes

Non-DEFLATE Algorithms

Archivers based on completely different compression algorithms often produce superior results.
Unfortunately, browsers only support gzip compression at the moment.
However, support for Brotli is constantly growing - but your browser doesn't support it.
Algorithm Program Parameters Size Compared To Best Zopfli
RAR (proprietary) RAR rar a -m5 -md64m -mc63:128t -mt1 3557 bytes -419 bytes (-10.54%)
Brotli (Wikipedia) brotli brotli -q 11 3655 bytes -321 bytes (-8.07%)
PPMd (Wikipedia) 7zip 7za a -mx=9 -m0=ppmd 3943 bytes -33 bytes (-0.83%)
Burrows-Wheeler transform (Wikipedia) bzip2 bzip2 -9 4014 bytes +38 bytes (+0.96%)
Zstandard (Wikipedia) zstd zstd -19 4014 bytes +38 bytes (+0.96%)
LZMA2 (Wikipedia) xz xz -9 4068 bytes +92 bytes (+2.31%)
ZPAQ (Wikipedia) zpaq zpaq -method 69 4418 bytes +442 bytes (+11.12%)

Detailled Analysis

I wrote a DEFLATE decoder in Javascript. Click the button below to start a client-side analysis of the smallest gzipped files (may take a second):


Notes: pigz is a fast open source multi-threaded implementation of gzip written by one of the original authors of gzip.
However, when using compression level 11, pigz actually switches to the slower Zopfli algorithm and isn't multi-threaded anymore.
KrzyMOD's extensions to Zopfli offer the highest level of configuration and is therefore used for my brute-force search.
Ken Silverman wrote the closed-source KZIP compression program and Jonathon Fowler ported it to Linux.
Defluff was created by Joachim Henke; DeflOpt is a tool by Ben Jos Walbeehm.

website made by Stephan Brumme in 2015 and still improving in 2024.
all timestamps are displayed in central european time. see my changelog.
no flash, not even images or external css files - and everything squeezed into a single html file.
which was handsomely compressed before releasing it into the wild internet - obviously.

please visit my homepage and my blog, too.
email: minime (at) stephan-brumme.com

All trademarks are property of their respective owners. You know, the boring legal stuff.