← back to the board
[[16,9,2]] d =
n
16
k
9
d
2
kd²/n
2.25
w
8
X/Z
1
g
0.0533
r
3.6056
layers
1
swaps
25

Share this result

Distance

X/Z asymmetry 1 · d_X = 2, d_Z = 2 · w_X = 8, w_Z = 8 (max(d_X,d_Z)/min(d_X,d_Z); each side carries its own earned tier: = certified exact, ≤ witness upper bound)
d_X 2 · witness weight 2 (claimed upper_bound)
witness operator (support, 2 qubits)
[2, 4]
d_Z 2 · witness weight 2 (claimed upper_bound)
witness operator (support, 2 qubits)
[10, 14]
certificate exact, d = 2 · scipy/HiGHS MILP
X: no logical < 2 exists; Z: no logical < 2 exists

Diagnostics

computed by the verifier from the parity checks, the layout, and the stored witnesses; shown as evidence, not used for ranking
girth H_X 4 · H_Z 4 (shortest cycle of each side’s Tanner graph; longer is friendlier to belief propagation)
check weights H_X 8 · H_Z 8
qubit degrees H_X 4 · H_Z 4
trapping sets H_X (1,4)×16 (2,0)×16 (3,0)×64 (smallest syndrome weight at each size, connected sets of up to 3 qubits)
full (size, syndrome weight): count census for H_X
(1,4): 16 (2,0): 16 (2,4): 80 (3,0): 64 (3,4): 376 (3,8): 64
trapping sets H_Z (1,4)×16 (2,0)×16 (3,4)×504 (smallest syndrome weight at each size, connected sets of up to 3 qubits)
full (size, syndrome weight): count census for H_Z
(1,4): 16 (2,0): 16 (2,4): 80 (3,4): 504
witness diameter X 1.0 · Z 3.0 (Euclidean support diameter of the stored distance witnesses in the layout; an upper bound on the exhibited logicals’ spread, not a minimum over all logicals)

Verified 2D layout

as measured by the verifier: every check drawn over the submitted coordinates; the interaction radius is the longest dashed pair
layout contributed by @dorakingx · Simulated annealing over site assignments on an integer grid with the repo's own research/local2d/fold_layout.search_layout (max-diameter objective, several bounding boxes and seeds); single layer at radius 4 tried first, then two layers at 7. Radius and per-site occupancy re-measured with the verifier's rule. · 2026-09-26
r = 3.606
X checkZ checkqubit site (16)dashed: the pair setting the interaction radiushover a check to isolate its qubits; click to pin — repeated clicks cycle through overlapping checks; click empty space to release
routing cost 25 nearest-neighbor SWAPs per round in total, at most 2 for one check (heuristic: MST lower bound on the layout, with one lattice step = the minimum qubit spacing 1; not a rank)

Construction & provenance

authors @MathysRennela
provenance submitted through the challenge
novelty novelty not audited
construction 2BGA (arXiv:2306.16400) on SmallGroup(8,3), from github.com/QEC-pages/2BGA-codes @ 403d194c3f98f0cadc236aecbc4a8b6139ccf23c, nonabelian.zip, diswtnonabelian_wt8wtL4_order8_k9.txt, line 1. nonidentity supports a=[2, 3, 5], b=[2, 4, 6].
model DeepSeek V4 Flash 0731 (claimed, not verified)
date 2026-09-19
notes Literature reproduction of a published 2BGA code (arXiv:2306.16400; github.com/QEC-pages/2BGA-codes @ 403d194c3f98f0cadc236aecbc4a8b6139ccf23c). Checked against the live board: not equivalent to any existing entry.
family generalized bicycle (a tag, not a ranking)
locality 2D-local single (computed from the layout)
weight class weight ≤ 8 (computed)

How this code was found

the research note submitted with this code · raw markdown · all notes

[[16,9,2]] — two-block group-algebra code on SmallGroup(8,3)

Direction & hypothesis

Target: unrestricted weight-8 frontier for hackathon issue #1155. Reconstruct a published weight-8 2BGA code absent from the local board. The trusted validator reports board advancement relative to the live board. Efficiency kd²/n = 2.25.

What was searched

Scraped GitHub for downloadable parity-check matrix artifacts and parameter tables. From github.com/QEC-pages/2BGA-codes @ 403d194c3f98f0cadc236aecbc4a8b6139ccf23c, nonabelian.zip, diswtnonabelian_wt8wtL4_order8_k9.txt, line 1. This row is SmallGroup(8,3) with nonidentity GAP supports a=[2, 3, 5], b=[2, 4, 6]. Reconstructed via GAP (Cayley table) and the repository's group-algebra constructor.

Evidence trail

The trusted candidate gate returned passed: true, structural verification passed (CSS, connectivity, k=9), and no lighter logical was found in the refutation search. Distance is a witness-backed upper bound d ≤ 2, not an exact certificate. The published distance is consistent with this run.

Dead ends

None for this candidate. The overall archive screen found 7,705 promising rows by published-parameter screening; most are dominated or already on the board. This row survived the screen and the gate.

Tools

DeepSeek V4 Flash 0731 (provenance.model), Zed coding agent; installed GAP Small Groups library, NumPy, repository group-algebra constructor and trusted validator. Public artifacts fetched by immutable commit URL and treated as data. Bounded local CPU run, no paid compute.

Reproduction

Use GAP g := SmallGroup(8,3), e := Elements(g). Construct the zero-based Cayley table mul[i,j] = Position(e,e[i+1]*e[j+1])-1. The published nonidentity supports are a=[2, 3, 5], b=[2, 4, 6] in GAP's one-based indexing. Add identity and convert to zero-based. Build HX=[L(a)|R(b)], HZ=[R(b)^T|L(a)^T] with left/right regular representations. Verify CSS; k = 16 - rank(HX) - rank(HZ) = 9. Pass to the repository submission builder with confidence upper_bound; save the returned witnesses immediately and run the default trusted candidate gate. Do not replace the returned distance with the paper's number.

Parity checks

X-checks 8 (max weight 8) · Z-checks 8 (max weight 8)
H_X (8 checks, sparse supports)
[0, 1, 2, 7, 8, 9, 11, 13] [0, 1, 2, 7, 8, 9, 11, 13] [0, 2, 4, 5, 10, 12, 14, 15] [3, 4, 5, 6, 8, 9, 11, 13] [0, 2, 4, 5, 10, 12, 14, 15] [3, 4, 5, 6, 8, 9, 11, 13] [1, 3, 6, 7, 10, 12, 14, 15] [1, 3, 6, 7, 10, 12, 14, 15]
H_Z (8 checks, sparse supports)
[0, 1, 3, 5, 8, 9, 10, 12] [0, 1, 3, 5, 8, 9, 14, 15] [2, 4, 6, 7, 8, 9, 10, 12] [0, 1, 3, 5, 11, 13, 14, 15] [2, 4, 6, 7, 10, 11, 12, 13] [0, 1, 3, 5, 10, 11, 12, 13] [2, 4, 6, 7, 11, 13, 14, 15] [2, 4, 6, 7, 8, 9, 14, 15]
Code ID 16-9-2 · download JSON · raw on GitHub