A(46, 6, 6)
75807581
+1
OPTES research · Mathematics
The OPTES search engine found 21 binary codes larger than the best published in the world table of constant-weight codes kept by Andries Brouwer. 18 of them are now the record in that table, credited to Mohammed Riyadh (OPTES).
In the background: real codewords from our A(46,8,8) code.
Take words of zeros and ones, each n positions long with exactly w ones. How many words can you collect at most, if any two of them must differ in at least d positions? That number is written A(n,d,w).
These numbers are basic to coding theory and error correction, and most of their exact values are still unknown. Each cell has a lower bound, the largest code anyone has found, and an upper bound that cannot be passed. Everything on this page raises a lower bound.
An example from our A(46,8,8) code: two of its 9704 words, 8 ones each, differing in exactly 8 positions (lit). That is the least the rule allows.
In Brouwer's table these cells now show our code as the record, with the reference «Riy: Mohammed Riyadh (OPTES)». Values as in the table update of 22 November 2026; the arrow goes from the previous record to the one credited now.
A(46, 6, 6)
75807581
+1
A(51, 6, 6)
1181911827
+8
We verified a larger code: 11832, not sent yet.
A(52, 6, 6)
1261912638
+19
We verified a larger code: 12647, not sent yet.
A(54, 6, 6)
1470814709
+1
A(48, 8, 7)
28892890
+1
A(49, 8, 7)
33123313
+1
A(50, 8, 7)
34703472
+2
A(51, 8, 7)
36283631
+3
We verified a larger code: 3632, not sent yet.
A(52, 8, 7)
38013804
+3
A(53, 8, 7)
42074208
+1
A(54, 8, 7)
47284729
+1
A(57, 8, 7)
60236025
+2
A(59, 8, 7)
64386441
+3
We verified a larger code: 6443, not sent yet.
A(40, 8, 8)
37523755
+3
A(42, 8, 8)
52045208
+4
We verified a larger code: 5214, not sent yet.
A(43, 8, 8)
62486254
+6
We verified a larger code: 6258, not sent yet.
A(45, 8, 8)
81378140
+3
We verified a larger code: 8148, not sent yet.
A(46, 8, 8)
91369474
+338
We verified a larger code: 9704, not sent yet.
These codes passed the same two checkers but have not been sent to the table's editor yet, so the table does not show them. Until they are added, their files below are the reference.
A(42, 6, 6)
51405142
+2
A(42, 8, 7)
13971398
+1
A(52, 10, 7)
423424
+1
A(42, 8, 8)
52085214
+6
A(43, 8, 8)
62546258
+4
A(45, 8, 8)
81408148
+8
A(46, 8, 8)
94749704
+230
A(51, 6, 6)
1182711832
+5
A(51, 8, 7)
36313632
+1
A(52, 6, 6)
1263812647
+9
A(59, 8, 7)
64416443
+2
Each file is plain text, one word per line. Check three things: every line has length n, every line has w ones, and any two lines differ in at least d positions. The parameters and the word count are in the file name, as in A_46_8_8_9704.txt.
| Cell | Words | Status | File | SHA-256 |
|---|---|---|---|---|
| A(40,8,8) | 3755 | credited | A_40_8_8_3755.txt 150 KB | 1f4f63501d0486e8… |
| A(42,6,6) | 5142 | not sent | A_42_6_6_5142.txt 216 KB | 0e89bc1ba029f5ab… |
| A(42,8,7) | 1398 | not sent | A_42_8_7_1398.txt 59 KB | 4a771f4c1a71fbca… |
| A(42,8,8) | 5214 | above credited | A_42_8_8_5214.txt 219 KB | bed08d0b2cdc7d0c… |
| A(43,8,8) | 6258 | above credited | A_43_8_8_6258.txt 269 KB | ac015ca5d8ee6abc… |
| A(45,8,8) | 8148 | above credited | A_45_8_8_8148.txt 366 KB | 77bb1edd5174633b… |
| A(46,6,6) | 7581 | credited | A_46_6_6_7581.txt 348 KB | 99582fda56c50019… |
| A(46,8,8) | 9704 | above credited | A_46_8_8_9704.txt 445 KB | e63efeb92b740517… |
| A(48,8,7) | 2890 | credited | A_48_8_7_2890.txt 138 KB | eb8312c21e7a7b6a… |
| A(49,8,7) | 3313 | credited | A_49_8_7_3313.txt 162 KB | d125f6873dfcab71… |
| A(50,8,7) | 3472 | credited | A_50_8_7_3472.txt 173 KB | a6859a7e76f127af… |
| A(51,6,6) | 11832 | above credited | A_51_6_6_11832.txt 601 KB | 44e323b9d4c5bb56… |
| A(51,8,7) | 3632 | above credited | A_51_8_7_3632.txt 184 KB | 430a5e5627fa184d… |
| A(52,6,6) | 12647 | above credited | A_52_6_6_12647.txt 655 KB | 6d2a15b9ebe59969… |
| A(52,8,7) | 3804 | credited | A_52_8_7_3804.txt 197 KB | 7946a82e4c40baa6… |
| A(52,10,7) | 424 | not sent | A_52_10_7_424.txt 22 KB | ad2de968a97dcb7f… |
| A(53,8,7) | 4208 | credited | A_53_8_7_4208.txt 222 KB | 23dc25638e01ad20… |
| A(54,6,6) | 14709 | credited | A_54_6_6_14709.txt 790 KB | e737d04a0614adb3… |
| A(54,8,7) | 4729 | credited | A_54_8_7_4729.txt 254 KB | 4430ccd2c970a162… |
| A(57,8,7) | 6025 | credited | A_57_8_7_6025.txt 341 KB | 5ab5f7918f3ac207… |
| A(59,8,7) | 6443 | above credited | A_59_8_7_6443.txt 378 KB | 5282355ac8a8e307… |
A short program that checks any file (the large ones take minutes):
# python3 check.py A_46_8_8_9704.txt
import itertools, os, re, sys
f = sys.argv[1]
n, d, w, size = map(int, re.findall(r"\d+", os.path.basename(f))[-4:])
words = [l.strip() for l in open(f) if l.strip()]
assert len(words) == size == len(set(words))
assert all(len(x) == n and x.count("1") == w and set(x) <= {"0", "1"} for x in words)
v = [int(x, 2) for x in words]
m = min(bin(a ^ b).count("1") for a, b in itertools.combinations(v, 2))
print(size, "words, smallest distance", m, "OK" if m >= d else "FAIL")
Before publication every file passed two independent checkers: the first measures the difference between every two words, the second makes sure no w − d/2 + 1 positions are all ones in two different words.