Puzzle swarm benchmark

Judge-verified results from the Siva Goldbug and Minerva Lux solving runs.

Snapshot captured Aug 24, 2026, 7:13 PM PDT. Running figures continue changing.

Overview

Goldbug solves
4 / 6
4 solved · 2 unsolved
Minerva Lux
10 / 15
Original swarm finished
Focused swarms
2
Puzzles 4 and 5 · xhigh
Observed tokens
≥ 43.06B
Some early tracking incomplete

Three Goldbug challenges were tested under open-web, leak-blocked, and hint-blocked conditions. The separate Minerva Lux swarm used solstice-alpha with high reasoning across all 15 puzzles; dedicated follow-up swarms now use solstice-alpha with xhigh reasoning on the two unresolved targets.

Goldbug puzzle gallery

Original challenge assets and the outcome of each valid solving run.

Dress puzzle source image: Lace strips, knots, colored beads, and an encoded dress message.

Dress 1 solved · 1 unsolved

Lace strips, knots, colored beads, and an encoded dress message. Solved with an official organizer clue; the separate hint-blind run failed.

Accepted answer: INCONSISTENTKNOTS

Cubes puzzle source image: Colored puzzle cubes whose hidden faces and perspective reveal letters.

Cubes 3 solved · 0 unsolved

Colored puzzle cubes whose hidden faces and perspective reveal letters. Three clean solves. Every successful route eventually used an official clue.

Accepted answer: LACKOFDUST

Ducks puzzle source image: Seven duck drawings with embedded cryptographic or steganographic details.

Ducks 0 solved · 1 unsolved

Seven duck drawings with embedded cryptographic or steganographic details. The valid leak-blocked run stopped without solving after 102 guesses.

Result: Never solved.

Verified Goldbug run timeline

Black denotes an accepted solution; gray denotes an unsolved run.

Goldbug puzzle runs by elapsed time and verified outcomeSix runs spanning Aug 22, 2:44 AM through Aug 22, 8:07 PM Pacific time.Aug 22, 2:44 AMAug 22, 7:05 AMAug 22, 11:26 AMAug 22, 3:46 PMAug 22, 8:07 PMCubes Open webCubes, Open web: Solved, 2h 02m2h 02mDucks Blocked leaksDucks, Blocked leaks: Unsolved, 6h 07m6h 07mDress Blocked leaksDress, Blocked leaks: Solved, 1h 19m1h 19mCubes Blocked leaksCubes, Blocked leaks: Solved, 2h 33m2h 33mCubes Hints blockedCubes, Hints blocked: Solved, 5h 25m5h 25mDress Hints blockedDress, Hints blocked: Unsolved, 4h 19m4h 19m

All run results

Six Goldbug benchmarks, one complete Minerva Lux suite, and two dedicated Minerva Lux follow-ups
RunJudge resultHintsDurationTokensGuesses
Cubes
Open web
SolvedOfficial clue2h 02m67.7M†5
Ducks
Blocked leaks
UnsolvedPublic clues6h 07m146.9M‡102
Dress
Blocked leaks
SolvedOfficial clue1h 19m7.98M†2
Cubes
Blocked leaks
SolvedOfficial clue2h 33m7
Cubes
Hints blocked
SolvedInjected later5h 25m2.45B‡6
Dress
Hints blocked
UnsolvedInjected, unused4h 19m1.69B‡20
Minerva Lux
15-puzzle swarm
10/15 solvedUnrestricted28h 18m36.86B191
Puzzle 4
Focused continuation
RunningPrior research46m889.3M34
Puzzle 5
Focused continuation
RunningPrior research46m951.7M93

† Main-agent tokens only. ‡ Partial swarm accounting. — No token counter survived. Focused-run durations describe the current xhigh continuation; their token totals include earlier activity.

Minerva Lux puzzle outcomes

Every row shows the actual source artwork, judge-verified answer when solved, required answer length, and submission count from the completed 15-puzzle swarm.

Private completed Minerva Lux monitor showing 10 of 15 puzzles solved, 191 attempts, and full token accounting
Completed full-suite monitor captured without exposing its private address.
10 solved; 5 unsolved; 191 total attempts in the original swarm
ArtworkPuzzleJudge resultAccepted answerAttempts
Full original artwork for Minerva Lux puzzle 1Puzzle 1
24 characters
Solvedcontrol alternate delete2
Full original artwork for Minerva Lux puzzle 2Puzzle 2
9 characters
Solved1465112083
Full original artwork for Minerva Lux puzzle 3Puzzle 3
9 characters
Solved2327925607
Full original artwork for Minerva Lux puzzle 4Puzzle 4
9 characters
Unsolved
Focused run active
34
Full original artwork for Minerva Lux puzzle 5Puzzle 5
15 characters
Unsolved
Focused run active
53
Full original artwork for Minerva Lux puzzle 6Puzzle 6
52 characters
Solvedthe soul becomes dyed with the color of its thoughts2
Full original artwork for Minerva Lux puzzle 7Puzzle 7
31 characters
Unsolved4
Full original artwork for Minerva Lux puzzle 8Puzzle 8
48 characters
Solvedlet the flood come and wash away the work of man2
Full original artwork for Minerva Lux puzzle 9Puzzle 9
11 characters
Unsolved40
Full original artwork for Minerva Lux puzzle 10Puzzle 10
11 characters
Solvedoppositorum3
Full original artwork for Minerva Lux puzzle 11Puzzle 11
14 characters
Solveda digital rain3
Full original artwork for Minerva Lux puzzle 12Puzzle 12
19 characters
Solvedfind form from flux3
Full original artwork for Minerva Lux puzzle 13Puzzle 13
22 characters
Solvedwalk like an eggyptian4
Full original artwork for Minerva Lux puzzle 14Puzzle 14
28 characters
Solveda water drop hollows a stone1
Full original artwork for Minerva Lux puzzle 15Puzzle 15
20 characters
Unsolved30

View the official Minerva Lux challenge

Dedicated Puzzle 4 and Puzzle 5 investigations

Each continuation has its own isolated solving container, judge, audit trail, token counter, and optional agent swarm. Dashboard snapshots are embedded instead of linking to private infrastructure.

Puzzle 4 Running

Original image for the focused Minerva Lux puzzle 4 investigation
Configuration
solstice-alpha · xhigh reasoning
Fresh attempts
34 new · 33 inherited rejections
Observed tokens
889,295,141 across 259 sessions
Latest candidate
363050040 · incorrect
Current segment
46m since the xhigh restart; cumulative token totals include the earlier paused segment.
Private Puzzle 4 monitoring dashboard showing live status, attempts, token usage, and recent agent activity
Private monitor captured at publication time; no internal endpoint is exposed.

Puzzle 5 Running

Original image for the focused Minerva Lux puzzle 5 investigation
Configuration
solstice-alpha · xhigh reasoning
Fresh attempts
93 new · 61 inherited rejections
Observed tokens
951,717,192 across 311 sessions
Latest candidate
RESISTOR BRIDGE · incorrect
Current segment
46m since the xhigh restart; cumulative token totals include the earlier paused segment.
Private Puzzle 5 monitoring dashboard showing live status, attempts, token usage, and recent agent activity
Private monitor captured at publication time; no internal endpoint is exposed.

Run notes and methodology

  • Cubes · Open web: Used the allowed organizer perspective clue and derived letters from solved cube fronts and undersides.
  • Ducks · Blocked leaks: No forbidden source observed. Submitted 102 candidates and stopped without solving.
  • Dress · Blocked leaks: Used the allowed organizer clue about U-110 and June 1941, reconstructed the dress cipher, and solved on the second submission.
  • Cubes · Blocked leaks: Used allowed organizer clues and correctly decoded LACK / OF / DUST. The process reported failure after the judge had accepted the answer.
  • Cubes · Hints blocked: Initially hint-blind. Official clues were injected at 01:54 UTC; the resumed swarm solved at 03:07 UTC. Its saved result incorrectly remains unsolved.
  • Dress · Hints blocked: Hint-blind for most of the run. A hint file arrived seven minutes before failure, with no evidence that the solver read it.
  • Minerva Lux · 15-puzzle swarm: Original swarm completed with 10 judge-confirmed solutions. Its saved summary reported 10 and cannot be treated as authoritative.
  • Puzzle 4 · Focused continuation: Dedicated single-puzzle xhigh-reasoning swarm. Inherited 33 historical rejections and has made 34 new attempts.
  • Puzzle 5 · Focused continuation: Dedicated single-puzzle xhigh-reasoning swarm. Inherited 61 historical rejections and has made 93 new attempts.
Outcome source
Persisted local judge ledgers; accepted answers override stopped, failed, or stale process summaries.
Hint interpretation
Official organizer clues count as hints. Injected clues are distinguished from clues never read.
Token accounting
At least 43,060,898,186 observed tokens; early main-only and partial counters mean the true total is higher.
Audit data
92.4 GB of persisted event logs across included runs, measured at capture time.
Included runs
Only valid benchmark runs with a functioning environment and trustworthy solving conditions.