Choose a version:
32% The original file has 45489 bytes (44.4k) and is available from the project website.
There you can find the official minified version, too, which brings down the size to 14682 bytes (14.3k, 32%).

After GZIP compression these minified files vary in size:
Boot
  5695 bytes (5.6k)
CDN, direct link
jsdelivr
  5695 bytes (5.6k)
CDN, direct link
gzip -6 (default)
  5087 bytes (5.0k)
local copy
Baidu
  5079 bytes (5.0k)
CDN, direct link
gzip -9
  5078 bytes (5.0k)
local copy
cdnjs
  5074 bytes (5.0k)
CDN, direct link
unpkg
  5074 bytes (5.0k)
CDN, direct link
cdnhttps
  5063 bytes (4.9k)
CDN, direct link
Yandex
  5054 bytes (4.9k)
CDN, direct link
7zip -mx=9 -tgzip
  4950 bytes (4.8k)
local copy
libdeflate -12
  4945 bytes (4.8k)
local copy
kzip -s0 -rn -b1
  4939 bytes (4.8k)
local copy
pigz -11 -n
  4928 bytes (4.8k)
local copy
Zopfli
  4925 bytes (4.8k)
local copy

perma-link to the smallest file on my server:
http://minime.stephan-brumme.com/files/underscorejs/underscore-1.6.0.min.js

You will automatically get the smallest UnderscoreJS 1.6.0 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 129 bytes by using my UnderscoreJS 1.6.0 Zopfli version instead of the best available CDN (2.62% smaller than Yandex, 4925 vs. 5054 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 --i10000 --mb8 --mls8 --bsr11 --lazy --ohh

(found September 19, 2015)
Description Value Parameter
iterations 10000  --i10000
maximum blocks 8  --mb8
maximum length score 8  --mls8
block splitting recursion 11  --bsr11
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/1.6.0/underscore-min.js --location | md5sum
dd9663be9a71f3570bc35f0edba28712  -
curl --silent --compressed http://minime.stephan-brumme.com/files/underscorejs/underscore-1.6.0.min.zopfli.js.gz | md5sum
dd9663be9a71f3570bc35f0edba28712  -

SHA1:
curl --silent --compressed https://raw.githubusercontent.com/jashkenas/underscore/1.6.0/underscore-min.js --location | sha1sum
08092373bfbdb7924dd01795779e171e66f038cd  -
curl --silent --compressed http://minime.stephan-brumme.com/files/underscorejs/underscore-1.6.0.min.zopfli.js.gz | sha1sum
08092373bfbdb7924dd01795779e171e66f038cd  -

These CDNs send you the original file:
CDN Size (compressed) MD5 (uncompressed) Timestamp
Boot 5695 bytes dd9663be9a71f3570bc35f0edba28712 (invalid)
jsdelivr 5695 bytes dd9663be9a71f3570bc35f0edba28712 (invalid)
cdnjs 5074 bytes dd9663be9a71f3570bc35f0edba28712 (invalid)
unpkg 5074 bytes dd9663be9a71f3570bc35f0edba28712 July 11, 2016 @ 16:49
cdnhttps 5063 bytes dd9663be9a71f3570bc35f0edba28712 December 24, 2015 @ 07:36
Yandex 5054 bytes dd9663be9a71f3570bc35f0edba28712 (invalid)

And some CDNs send you a different file:
CDN Size (compressed) MD5 (uncompressed) Comment / Diff Timestamp
Baidu 5079 bytes 5a9d9c80ed25f4239f063070289d2a74 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:

1.8.3, 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
4925 bytes -1 byte zopfli --i10000 --mls8 --bsr11 --lazy --ohh September 19, 2015 @ 08:23
4926 bytes -1 byte zopfli --i10000 --mls1024 --bsr14 --lazy --ohh September 8, 2015 @ 12:15
4927 bytes -1 byte zopfli --i1000 --mls512 --bsr10 --lazy --ohh September 8, 2015 @ 10:40
4928 bytes zopfli --i100 --mls4096 --bsr12 --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 11, 2016 @ 16:49.

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 or 100,000.
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
4928 4928 4928 4928 4928 4928 4928 4929 4927 4929 4927 4927 4929 4929 4929
4927 4927 4926 4935 4929 4926 4928 4928 4928 4931 4930 4933 4928 4928 4928
4927 4927 4928 4928 4929 4929 4929 4929 4928 4928 4927 4929 4930 4929 4929
4930 4930 4926 4927 4929 4927 4926 4928 4928 4928 4928 4927 4929 4931 4931
4928 4928 4928 4929 4928 4927 4928 4927 4929 4930 4926 4927 4931 4931 4931
4927 4927 4926 4927 4928 4927 4928 4928 4928 4934 4926 4926 4930 4928 4928
4927 4927 4925 4928 4927 4927 4927 4929 4926 4930 4927 4926 4932 4928 4928
4929 4929 4925 4928 4928 4928 4928 4926 4928 4932 4926 4927 4927 4930 4930
4928 4928 4926 4926 4926 4927 4926 4928 4931 4931 4926 4926 4927 4931 4931
4927 4927 4926 4927 4928 4926 4927 4929 4928 4928 4927 4929 4933 4931 4931
4928 4928 4929 4927 4929 4926 4926 4928 4928 4926 4927 4928 4928 4931 4931
4929 4929 4926 4928 4928 4928 4928 4928 4927 4930 4927 4927 4931 4931 4931
4926 4926 4927 4926 4926 4928 4926 4928 4927 4926 4926 4926 4928 4931 4931
4927 4927 4925 4929 4928 4926 4928 4928 4927 4928 4927 4927 4928 4931 4931
4927 4927 4926 4927 4927 4926 4926 4926 4927 4930 4927 4929 4931 4928 4928
4927 4927 4926 4926 4927 4927 4926 4928 4927 4927 4926 4927 4931 4931 4931
4927 4927 4926 4926 4926 4927 4926 4928 4927 4931 4927 4928 4928 4931 4931
4926 4926 4926 4928 4928 4928 4926 4928 4927 4930 4927 4927 4931 4931 4931
4927 4927 4927 4926 4926 4932 4926 4926 4927 4927 4927 4929 4931 4931 4931
4927 4927 4925 4926 4928 4933 4926 4928 4927 4931 4927 4927 4931 4928 4928
4927 4927 4927 4927 4928 4928 4926 4929 4927 4931 4926 4926 4931 4928 4928
4927 4927 4926 4926 4928 4927 4927 4933 4927 4931 4931 4928 4931 4931 4931
4928 4928 4927 4927 4928 4926 4928 4929 4927 4931 4927 4927 4928 4928 4928

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 4928 bytes 100%
1,000 4926 bytes -2 bytes 100%
10,000 4925 bytes -1 byte 100%
100,000 4925 bytes 16.81%
1,000,000
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
4940 bytes +15 bytes (+0.30%) +1 byte
4939 bytes +14 bytes (+0.28%)
4961 bytes +36 bytes (+0.73%) +22 bytes
4959 bytes +34 bytes (+0.69%) +20 bytes
4987 bytes +62 bytes (+1.26%) +48 bytes
5020 bytes +95 bytes (+1.93%) +81 bytes
5055 bytes +130 bytes (+2.64%) +116 bytes
5071 bytes +146 bytes (+2.96%) +132 bytes
5103 bytes +178 bytes (+3.61%) +164 bytes

Non-DEFLATE Algorithms

Archivers based on completely different compression algorithms often produce superior results.
Unfortunately, browsers only support gzip compression at the moment.
Algorithm Program Parameters Size Compared To Best Zopfli
RAR (proprietary) RAR rar a -m5 -md64m -mc63:128t -mt1 4386 bytes -539 bytes (-10.94%)
Brotli (Wikipedia) brotli brotli -q 11 4593 bytes -332 bytes (-6.74%)
PPMd (Wikipedia) 7zip 7za a -mx=9 -m0=ppmd 4947 bytes +22 bytes (+0.45%)
ZSTD (Wikipedia) zstd zstd -19 4947 bytes +22 bytes (+0.45%)
Burrows-Wheeler transform (Wikipedia) bzip2 bzip2 -9 4975 bytes +50 bytes (+1.02%)
LZMA2 (Wikipedia) xz xz -9 5000 bytes +75 bytes (+1.52%)
ZPAQ (Wikipedia) zpaq zpaq -method 69 5218 bytes +293 bytes (+5.95%)

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 2017.
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