Choose a version:
32% The original file has 52374 bytes (51.1k) and is available from the project website.
There you can find the official minified version, too, which brings down the size to 16523 bytes (16.1k, 32%).

After GZIP compression these minified files vary in size:
Boot
  6561 bytes (6.4k)
 CDN
jsdelivr
  6561 bytes (6.4k)
 CDN
cdnjs
  5850 bytes (5.7k)
 CDN
gzip -6 (default)
  5846 bytes (5.7k)
 local copy
gzip -9
  5829 bytes (5.7k)
 local copy
cdnhttps
  5822 bytes (5.7k)
 CDN
unpkg
  5815 bytes (5.7k)
 CDN
Yandex
  5807 bytes (5.7k)
 CDN
libdeflate -12
  5658 bytes (5.5k)
 local copy
7zip -mx=9 -tgzip
  5653 bytes (5.5k)
 local copy
kzip -s0 -rn -b1
  5643 bytes (5.5k)
 local copy
pigz -11 -n
  5641 bytes (5.5k)
 local copy
Zopfli
  5629 bytes (5.5k)
 local copy

perma-link to the smallest file on my server:
http://minime.stephan-brumme.com/files/underscorejs/underscore-1.8.2.min.js (or via HTTPS)

You will automatically get the smallest UnderscoreJS 1.8.2 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 178 bytes by using my UnderscoreJS 1.8.2 Zopfli version instead of the best available CDN (3.16% smaller than Yandex, 5629 vs. 5807 bytes):
Download
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 --i100000 --mb8 --mls1024 --bsr16 --lazy --ohh

(found September 8, 2015)
Description Value Parameter
iterations 100000  --i100000
maximum blocks 8  --mb8
maximum length score 1024  --mls1024
block splitting recursion 16  --bsr16
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/1.8.2/underscore-min.js --location | md5sum
b0e9839a6bb6f12774494fa30c06bcdf  -
curl --silent --compressed https://minime.stephan-brumme.com/files/underscorejs/underscore-1.8.2.min.zopfli.js.gz | md5sum
b0e9839a6bb6f12774494fa30c06bcdf  -

SHA1:
curl --silent --compressed https://raw.githubusercontent.com/jashkenas/underscore/1.8.2/underscore-min.js --location | sha1sum
e686abc5c68dfaf7a14ce3dc68e33703750bc901  -
curl --silent --compressed https://minime.stephan-brumme.com/files/underscorejs/underscore-1.8.2.min.zopfli.js.gz | sha1sum
e686abc5c68dfaf7a14ce3dc68e33703750bc901  -

All listed CDNs deliver identical contents:
CDN Size (compressed) MD5 (uncompressed) Timestamp
Boot 6561 bytes b0e9839a6bb6f12774494fa30c06bcdf (invalid)
jsdelivr 6561 bytes b0e9839a6bb6f12774494fa30c06bcdf (invalid)
cdnjs 5850 bytes b0e9839a6bb6f12774494fa30c06bcdf (invalid)
cdnhttps 5822 bytes b0e9839a6bb6f12774494fa30c06bcdf December 24, 2015 @ 07:36
unpkg 5815 bytes b0e9839a6bb6f12774494fa30c06bcdf July 11, 2016 @ 16:48
Yandex 5807 bytes b0e9839a6bb6f12774494fa30c06bcdf December 22, 2015 @ 17:09

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

Other Versions

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

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
5629 bytes -1 byte zopfli --i100000 --mls1024 --bsr16 --lazy --ohh September 8, 2015 @ 18:51
5630 bytes -1 byte zopfli --i10000 --mls1024 --bsr16 --lazy --ohh September 8, 2015 @ 11:09
5631 bytes -1 byte zopfli --i1000 --mls1024 --bsr16 --lazy --ohh September 8, 2015 @ 10:38
5632 bytes – zopfli --i100 --mls8192 --bsr8 --lazy --ohh September 8, 2015 @ 10:34

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 @ 16:48.

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
2 5640 5633 5640 5632 5640 5636 5635 5631 5630 5640 5640 5639 5633 5635 5635
4 5647 5647 5647 5635 5640 5640 5636 5636 5631 5640 5640 5637 5638 5638 5638
6 5646 5632 5646 5648 5640 5636 5636 5635 5630 5630 5640 5637 5637 5638 5638
7 5635 5635 5647 5634 5640 5640 5636 5630 5631 5632 5640 5638 5631 5638 5638
8 5633 5633 5636 5634 5640 5635 5636 5631 5631 5630 5640 5637 5630 5632 5632
9 5647 5647 5647 5635 5648 5640 5636 5635 5631 5640 5640 5634 5637 5638 5638
10 5648 5647 5636 5634 5647 5635 5629 5636 5630 5640 5640 5638 5637 5638 5638
11 5633 5633 5636 5634 5640 5640 5636 5631 5631 5630 5640 5635 5635 5638 5638
12 5646 5646 5647 5634 5640 5635 5636 5636 5635 5640 5640 5635 5630 5632 5632
13 5648 5647 5647 5634 5640 5635 5636 5630 5630 5630 5640 5636 5629 5637 5637
14 5647 5646 5646 5635 5640 5636 5636 5631 5631 5631 5640 5637 5637 5638 5638
15 5634 5634 5646 5634 5640 5635 5636 5630 5631 5632 5640 5638 5638 5638 5638
16 5646 5646 5647 5634 5640 5636 5636 5637 5631 5629 5640 5637 5629 5638 5638
17 5647 5646 5636 5634 5640 5635 5636 5630 5636 5629 5640 5638 5631 5632 5632
18 5634 5634 5647 5634 5648 5635 5636 5630 5633 5629 5640 5639 5631 5637 5637
19 5635 5634 5636 5635 5640 5636 5636 5634 5636 5632 5640 5636 5638 5638 5638
20 5646 5646 5647 5634 5647 5631 5636 5631 5631 5630 5640 5638 5638 5638 5638
21 5647 5646 5636 5647 5640 5635 5636 5636 5631 5630 5640 5638 5638 5638 5638
22 5646 5646 5647 5634 5640 5635 5636 5630 5632 5632 5640 5636 5638 5638 5638
23 5634 5634 5647 5634 5647 5635 5636 5634 5637 5640 5640 5638 5631 5638 5638
25 5646 5646 5646 5634 5640 5635 5636 5636 5632 5631 5640 5636 5638 5637 5637
30 5645 5647 5647 5635 5647 5635 5636 5637 5636 5631 5640 5638 5638 5638 5638
40 5647 5647 5646 5635 5648 5631 5636 5637 5632 5632 5640 5637 5636 5638 5638

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 5632 bytes – 100%
1,000 5631 bytes -1 byte 100%
10,000 5629 bytes -2 bytes 100%
100,000 5629 bytes – 16.23%
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
0 5643 bytes +14 bytes (+0.25%) –
1 5643 bytes +14 bytes (+0.25%) –
2 5664 bytes +35 bytes (+0.62%) +21 bytes
3 5698 bytes +69 bytes (+1.23%) +55 bytes
4 5721 bytes +92 bytes (+1.63%) +78 bytes
5 5729 bytes +100 bytes (+1.78%) +86 bytes
6 5767 bytes +138 bytes (+2.45%) +124 bytes
7 5798 bytes +169 bytes (+3.00%) +155 bytes
8 5830 bytes +201 bytes (+3.57%) +187 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 4955 bytes -674 bytes (-11.97%)
PPMd (Wikipedia) 7zip 7za a -mx=9 -m0=ppmd 5192 bytes -437 bytes (-7.76%)
Brotli (Wikipedia) brotli brotli -q 11 5262 bytes -367 bytes (-6.52%)
Burrows-Wheeler transform (Wikipedia) bzip2 bzip2 -9 5553 bytes -76 bytes (-1.35%)
Zstandard (Wikipedia) zstd zstd -19 5649 bytes +20 bytes (+0.36%)
LZMA2 (Wikipedia) xz xz -9 5696 bytes +67 bytes (+1.19%)
ZPAQ (Wikipedia) zpaq zpaq -method 69 5751 bytes +122 bytes (+2.17%)

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):
Load and Parse Data


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 2018.
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