Independent Analysis of AI Agents Playing Stronghold Crusader

Each model runs a medieval village for 33 to 45 game minutes on each of three maps. It sees the game through screenshots and a reading of the town, and acts with the mouse and keyboard. We score the economy it builds against doing nothing and against a game a person played on the same map.

Crusader Bench Score

Only models with a full-budget score on all 3 maps are compared here, and each one’s score is the mean of its map scores.

From the Interviews

“My biggest mistake was being cautious about expansion while being insufficiently rigorous about productivity.”

GPT-6.1 SolIts interview, after the gameThe game
  1. v1.6.3 · 13 October 2026Fixes for what the seven 1.6.2 runs on The Sinkhole showed: none made a run invalid, but they cost the models turns and game time.
  2. v1.6.2 · 12 October 2026The armory's storage, found missing in GLM 5.3 Flash's first run on The Sinkhole: its armory filled up with 14 workshops behind it, and nothing told it.
  3. v1.6.1 · 12 October 2026A new map, and one line of the rules reworded.
  4. v1.6.0 · 12 October 2026Weapons and armour join the economy, and the fixes listed on the roadmap from the 1.5 runs.
  5. v1.5.6 · 12 October 2026Prompt fixes before rerunning every model on Large Barren Desert.
  6. v1.5.5 · 12 October 2026Placement fixes after Kimi K3's Large Barren Desert run, whose first Marketplace was built but reported refused.
  7. v1.5.4 · 11 October 2026A patch at the maintainers' choice, though what the model is shown changed on every map: a reminder after long waits, and trees told apart from cacti.
  8. v1.5.3 · 11 October 2026Comparable with 1.5.0 to 1.5.2 on Oasis by the Sea; the Large Barren Desert rules changed.
  9. v1.5.2 · 11 October 2026Comparable with 1.5.0 and 1.5.1: nothing an Oasis by the Sea run sees changed.
  10. v1.5.1 · 11 October 2026Comparable with 1.5.0: runs on the profiles that 1.5.0 accepted send and play the same.
  11. v1.5.0 · 11 October 2026Comparable within itself only: market_trade sells in the game's own steps and never more than asked, and its description says how the game trades.
  12. v1.4.0 · 11 October 2026Comparable within itself only: the observation stats leave out goods with none stored, and the tool definitions and the preparation request are sent in fewer tokens.
  13. v1.3.0 · 10 October 2026Comparable within itself only: build_structure reports and retries more cases and takes map tiles, the host sends text instead of a screenshot after a turn without game_action, inspect_building or a pixel placement, screenshots stay in the conversation until compaction, every failure names its remedy, set_tax reads a live figure, and three little-used tools were removed.
  14. v1.2.1 · 10 October 2026Comparable with 1.2.0: a fix to what the agent is shown of its popularity.
  15. v1.2.0 · 10 October 2026Not comparable with 1.1.4: the agent is told that population and popularity count, sees more of its settlement, has new tools, and no longer carries a playbook between episodes.
  16. v1.1.4 · 10 October 2026Comparable with 1.1.3: a fix for runs the memory guard ended.
  17. v1.1.3 · 9 October 2026Comparable with 1.1.2: runs play and score the same.
  18. v1.1.2 · 9 October 2026Comparable with 1.1.1: runs play and score the same.
  19. v1.1.1 · 9 October 2026Comparable with 1.1.0: a placement fix.
  20. v1.1.0 · 9 October 2026Not comparable with 1.0.0: the agent sees storage readings and a changed prompt.
  21. v1.0.0 · 8 October 2026The first versioned benchmark, released as v1.

Score Breakdown 3 maps, Free Build

Models with a full-budget score on every map, scored by the mean of their map scores. Each map is scored from 0 for doing nothing to 100 for the person's game there, so a map with more room to grow does not count for more.

Mean over every map · 0 is doing nothing, 100 the person's game · Higher is better

010203040506050.446.636.834.227.623.920.4
Claude Opus 5.5 (medium)
GPT-6.1 Sol (medium)
GPT-6 Luna (medium)
Claude Sonnet 5.5 (medium)
GLM 5.3 Flash (medium)
Kimi K3 (high)
Claude Haiku 5.5 (medium)

Crusader Bench · crusbench.com

The score is 100 × (0.8 × net worth growth + 0.1 × population growth + 0.1 × popularity ÷ 100). Growth is measured on each map from doing nothing, 0, to the person’s game, 1. Population growth and popularity stop at the person’s level; net worth growth can run from −1 to 2. A map’s score is the mean of a model’s full-budget episodes on it, and the score over every map is the mean of its map scores.

Net worth is gold plus every stored good valued at the game’s fixed marketplace sell price: wood 1, stone 7, iron 23, bread, cheese, meat and apples 4, wheat and hops 8, flour, ale and pitch 10. Scoring in full

Minute by Minute

Economy, settlement, stock and agent metrics, sampled once per game minute.

Every game on every map, each model's colour lighter on the later map · choose one map for its baseline and the person's game

Economy
Settlement
Stock
Agent
Net worth · gold, by game minute
05,00010,00015,00020,00025,000051015202530354045 min
  • Claude Opus 5.5 · Oasis
  • Claude Opus 5.5 · Desert
  • Claude Opus 5.5 · The Sinkhole
  • GPT-6.1 Sol · Oasis
  • GPT-6.1 Sol · Desert
  • GPT-6.1 Sol · The Sinkhole
  • GPT-6 Luna · Oasis
  • GPT-6 Luna · Desert
  • GPT-6 Luna · The Sinkhole
  • Claude Sonnet 5.5 · Oasis
  • Claude Sonnet 5.5 · Desert
  • Claude Sonnet 5.5 · The Sinkhole
  • GLM 5.3 Flash · Oasis
  • GLM 5.3 Flash · Desert
  • GLM 5.3 Flash · The Sinkhole
  • Kimi K3 · Oasis
  • Kimi K3 · Desert
  • Kimi K3 · The Sinkhole
  • Claude Haiku 5.5 · Oasis
  • Claude Haiku 5.5 · Desert
  • Claude Haiku 5.5 · The Sinkhole
Show the data as a table
MinuteClaude Opus 5.5 · OasisClaude Opus 5.5 · DesertClaude Opus 5.5 · The SinkholeGPT-6.1 Sol · OasisGPT-6.1 Sol · DesertGPT-6.1 Sol · The SinkholeGPT-6 Luna · OasisGPT-6 Luna · DesertGPT-6 Luna · The SinkholeClaude Sonnet 5.5 · OasisClaude Sonnet 5.5 · DesertClaude Sonnet 5.5 · The SinkholeGLM 5.3 Flash · OasisGLM 5.3 Flash · DesertGLM 5.3 Flash · The SinkholeKimi K3 · OasisKimi K3 · DesertKimi K3 · The SinkholeClaude Haiku 5.5 · OasisClaude Haiku 5.5 · DesertClaude Haiku 5.5 · The Sinkhole
01,2954,5462,4151,2804,5462,4151,2804,5612,4151,2754,5012,4151,2654,5462,4151,2754,5312,4151,2704,5612,415
11,3846,1962,3781,3896,5342,3711,3676,4982,3791,0905,4092,3821,3715,4032,3891,3716,4282,3861,3945,4392,389
21,0766,3892,0461,0556,4442,0511,2186,0462,2611,0485,9932,0991,1806,4522,3601,3516,3632,1711,3806,4962,339
39026,0321,9977546,0411,9531,2185,9511,9618386,2472,0181,0656,4052,1741,3505,6312,1051,3926,5032,331
49355,9922,0267045,9791,8931,1905,9191,9618436,4861,7591,0136,4511,9451,3505,2191,4321,3606,5252,329
51,0866,0222,0157045,3561,5809835,6421,9217476,3821,7221,0906,2991,9691,1595,2861,4321,1716,5632,296
61,1346,0871,9029015,3361,5391,0065,3631,3888385,9181,7399516,2991,9341,0925,3011,4231,1856,3702,159
71,1386,1411,7147245,2041,6331,0045,1531,2729725,9161,4419796,3262,1238964,6621,2211,2506,3702,155
81,1276,0651,2467245,2041,4011,1835,1191,2761,1995,9161,6887956,3202,1228554,9501,2211,3576,3682,175
91,2786,1371,3321,3595,4231,4011,3265,2071,0381,2536,1101,6887236,0412,0439074,9021,3471,3806,3262,036
101,4836,0341,2901,1455,3752,0801,4324,5061,1111,4096,1811,9921,0706,0412,0437594,4281,3471,4056,3261,744
111,7886,0971,4101,1304,5892,3361,4674,6451,2541,7586,1812,0121,0855,9772,3458094,2871,3471,4906,1521,756
121,8526,2221,6201,1304,6642,3791,6384,5211,1861,9696,2361,8641,1416,0112,2249033,9101,5651,6426,1521,781
132,4015,8761,3321,7364,4662,6791,9594,6881,3342,2455,9202,1401,2076,0111,8011,2733,9611,5651,6736,1521,550
142,3035,9711,6581,6984,4572,6791,9644,6971,3712,4605,9202,1851,5646,0111,8011,3793,6721,7751,7636,1971,573
152,6806,0171,4351,8464,6193,4781,9644,3401,3712,4605,9202,6181,6066,2871,8581,4313,6721,7752,0616,1971,644
162,2986,0501,5811,7844,7713,3342,5614,6621,8543,0435,9202,8491,7855,9721,7231,7764,3311,7162,4216,3101,737
172,5646,0221,8501,7844,0393,5062,5614,4782,1233,0435,9203,2701,9076,1271,7231,9474,0771,7822,5976,3101,881
182,7966,2371,9792,7464,0973,4343,0594,9532,1233,0435,9203,1252,2835,9091,7232,0854,0451,7822,7086,2161,939
192,9006,4022,1572,8524,0973,7913,2815,0092,1233,8516,0203,1252,2835,9151,9122,2244,4712,2863,0026,2162,017
203,2596,4022,4052,8524,0974,0083,2815,1533,0503,8516,0203,8312,2835,9151,7552,2824,2192,2862,9256,2912,072
213,5706,7162,4802,8524,9584,5113,7294,6853,3933,8516,0204,4703,0715,9151,7401,9544,3612,7042,9846,2912,072
223,9756,7162,6194,0595,2204,8613,4154,8103,5083,8516,0204,4703,0715,9151,7401,9544,3432,7043,0946,3231,878
234,3476,9243,0594,0595,4615,0863,4154,8393,6793,8516,0205,4373,6536,4671,9102,7524,4353,2723,2206,3231,955
244,8156,9243,0594,8335,6205,0864,0994,8393,9435,6276,1145,4373,6776,0282,0162,7524,7003,3853,3186,4482,095
254,7227,3063,9435,0165,6206,1444,2464,8394,2395,6276,1146,3864,2016,0282,0163,2454,9953,3853,5236,4482,147
265,5337,3064,3185,4575,6206,5064,2465,2384,5285,6276,1147,0574,1926,0282,0163,6155,0623,3853,6576,5062,190
276,0467,7014,3275,5305,7136,7324,8305,2384,5286,8836,1147,0574,1926,0282,8233,6555,0173,8143,9066,5062,190
286,5257,7014,5145,5305,8247,3464,8305,2384,5286,8836,1147,0574,6706,4612,7323,9685,0213,8144,2256,2672,328
297,2447,7014,5575,5305,8247,6934,8306,0365,6146,8836,0198,1314,6706,2172,6803,9685,4074,4284,2256,2672,234
307,8117,7015,0036,7225,8247,9294,8306,4935,8866,8836,0198,1315,1366,3522,6804,7445,3864,8374,7286,2672,234
318,6077,9916,0006,6976,5697,9296,3886,4936,0646,8836,0199,3625,1366,3522,6804,7445,7634,9514,8936,3122,305
329,2397,9916,6676,8356,5698,8836,3886,4936,2948,2616,0199,3625,1366,3523,1855,2605,5824,9514,9576,3122,305
339,2397,9916,9717,2177,2679,6266,3886,8656,2948,2616,0199,3626,1196,3523,2515,2605,4935,6864,9576,3122,305
3410,0548,7577,4587,4017,44110,3327,3487,3546,2948,6336,25210,7606,2006,6553,2515,9185,7415,6865,1136,4122,388
35——8,310——10,919——7,530——10,760——3,540——5,712——2,388
36——8,478——11,613——7,452——12,183——3,507——5,712——2,725
37——8,361——12,304——7,535——12,183——3,507——6,101——2,776
38——9,127——13,410——7,624——12,183——3,507——6,521——2,759
39——9,691——14,453——8,024——12,183——4,010——6,866——2,759
40——10,312——15,130——8,526——13,789——3,729——6,872——2,759
41——11,011——15,939——8,755——13,789——3,729——6,872——2,908
42——11,796——15,939——9,354——13,789——3,729——7,996——2,908
43——11,796——18,382——9,354——15,524——4,317——7,996——2,908
44——12,426——18,382——10,543——15,524——4,317——7,996——3,173
45——13,921——20,610——11,177——16,810——5,084——9,164——3,312

Crusader Bench · crusbench.com

Resources

Gold, stored goods and food at the end of each episode, on every map.

Final reading after the harness pauses the game · units unless stated · each model's best episode on each map

  • Oasis by the Sea
  • Large Barren Desert
  • The Sinkhole
  • Each model’s colour, lighter on each later map

Gold

in the treasury
05,00010,00015,00020,000589
GPT-6.1 Sol (medium)
Claude Sonnet 5.5 (medium)
GLM 5.3 Flash (medium)
Claude Haiku 5.5 (medium)
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
Kimi K3 (high)

Goods Value

gold at sell price
05,00010,00015,000
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
GPT-6.1 Sol (medium)
Claude Sonnet 5.5 (medium)
Kimi K3 (high)
Claude Haiku 5.5 (medium)
GLM 5.3 Flash (medium)

Iron

units · 23 gold each
0100200300400178993191531003352719215414196104721054554910291748
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
Claude Sonnet 5.5 (medium)
GPT-6.1 Sol (medium)
Kimi K3 (high)
Claude Haiku 5.5 (medium)
GLM 5.3 Flash (medium)

Stone

units · 7 gold each
02004006003365061901773141761125302019331414024050204214892255020
Claude Opus 5.5 (medium)
GPT-6.1 Sol (medium)
GPT-6 Luna (medium)
Kimi K3 (high)
Claude Haiku 5.5 (medium)
GLM 5.3 Flash (medium)
Claude Sonnet 5.5 (medium)

Wood

units · 1 gold each
02004006005131413083028716339391312569716924824153198017641057
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)
Claude Opus 5.5 (medium)
Kimi K3 (high)
GLM 5.3 Flash (medium)
GPT-6 Luna (medium)
GPT-6.1 Sol (medium)

Food

units in the granary
010020030042981526681992401761864859862331236149057
GPT-6.1 Sol (medium)
GPT-6 Luna (medium)
Claude Sonnet 5.5 (medium)
Claude Opus 5.5 (medium)
GLM 5.3 Flash (medium)
Kimi K3 (high)
Claude Haiku 5.5 (medium)

Crusader Bench · crusbench.com

Selling at the marketplace converts goods to gold at the price the score already uses, so it does not change net worth. It does free up the stockpile, and production stops when the stockpile is full.

Settlement

The people, the popularity and the buildings each model left at the end, read from the game after the final pause.

Final reading · the game sends no more peasants while 24 stand idle at the campfire

  • Oasis by the Sea
  • Large Barren Desert
  • The Sinkhole
  • Each model’s colour, lighter on each later map

Population

peasants
050100150200130881831075814786106108769491647078602783613237
GPT-6.1 Sol (medium)
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
Kimi K3 (high)
GLM 5.3 Flash (medium)
Claude Sonnet 5.5 (medium)
Claude Haiku 5.5 (medium)

Housing

places
01002003002109811413090194154130114138581629811412274114106824258
GLM 5.3 Flash (medium)
GPT-6.1 Sol (medium)
GPT-6 Luna (medium)
Claude Opus 5.5 (medium)
Kimi K3 (high)
Claude Sonnet 5.5 (medium)
Claude Haiku 5.5 (medium)

Popularity

0 to 100
0501001501006679906679815752825948696652842468675151
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)
GPT-6 Luna (medium)
Claude Opus 5.5 (medium)
Kimi K3 (high)
Claude Sonnet 5.5 (medium)
Claude Haiku 5.5 (medium)

Idle Peasants

waiting at the campfire
010203024242424242424192424241724241623112424424
GLM 5.3 Flash (medium)
Kimi K3 (high)
Claude Sonnet 5.5 (medium)
GPT-6 Luna (medium)
Claude Haiku 5.5 (medium)
GPT-6.1 Sol (medium)
Claude Opus 5.5 (medium)

Worker Efficiency

per cent
05010015010010010010010010010010010010010010010010010010010010090100100
Claude Opus 5.5 (medium)
GPT-6.1 Sol (medium)
GPT-6 Luna (medium)
Claude Sonnet 5.5 (medium)
GLM 5.3 Flash (medium)
Claude Haiku 5.5 (medium)
Kimi K3 (high)

Buildings

standing at the end
05010015020088831508250120638483666382594565422360431435
GPT-6.1 Sol (medium)
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
Kimi K3 (high)
GLM 5.3 Flash (medium)
Claude Sonnet 5.5 (medium)
Claude Haiku 5.5 (medium)

Crusader Bench · crusbench.com

Population and popularity make up 20% of the score, 10% each; population is measured as growth from doing nothing towards the person’s game. Worker efficiency rises with fear, from gallows, gibbets and the game’s other “bad things”, and falls with “good things” such as gardens.

Tokens & Cost

Tokens used and what the games cost across every map, and the score against that cost.

Every game on every map · cache reads, cache writes, uncached input and output

  • Cache reads
  • Cache writes
  • Uncached input
  • Output
  • Shades of each model’s colour
05.0M10M15M20M19M17M15M14M9.9M8.6M7.2M
GPT-6 Luna (medium)
GPT-6.1 Sol (medium)
Claude Opus 5.5 (medium)
Kimi K3 (high)
GLM 5.3 Flash (medium)
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)

Crusader Bench · crusbench.com

The game is paused while the model thinks, so token use does not affect how much game time it gets. Each turn that runs the game lasts at least 8 game seconds, which limits the number of requests in a game and therefore its cost. A cost is what OpenRouter billed; where an API reports tokens but no bill, such as Anthropic’s and Moonshot’s, it is the game’s tokens at the model’s list prices, marked ~. The interview after an episode is not in the game’s cost.

Telemetry

Turns, tool calls, building placements, errors and time for each episode, from its record.

Per episode · inference time is spent with the game paused · each model's best episode on each map

  • Oasis by the Sea
  • Large Barren Desert
  • The Sinkhole
  • Each model’s colour, lighter on each later map

Turns

model replies
05010015012714313411811913613495521016697896268763056482958
GPT-6 Luna (medium)
GPT-6.1 Sol (medium)
Kimi K3 (high)
Claude Opus 5.5 (medium)
GLM 5.3 Flash (medium)
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)

Tool Calls

per episode
010020030018417222219112817416313812512312113811710311813651999154104
GPT-6 Luna (medium)
Claude Opus 5.5 (medium)
Kimi K3 (high)
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)

Placement Success

placed / attempted
0%50%100%150%98%93%99%84%96%85%96%96%91%69%98%83%79%92%87%59%92%83%50%93%
GPT-6 Luna (medium)
Claude Haiku 5.5 (medium)
Kimi K3 (high)
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)
Claude Opus 5.5 (medium)
Claude Sonnet 5.5 (medium)

Tool Errors

per episode
05124311121121112111021
Claude Haiku 5.5 (medium)
Claude Opus 5.5 (medium)
GPT-6 Luna (medium)
Claude Sonnet 5.5 (medium)
Kimi K3 (high)
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)

Inference Time

minutes, game paused
010203040352625252932182219171311131011745535
Kimi K3 (high)
GPT-6 Luna (medium)
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)
Claude Opus 5.5 (medium)
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)

Wall-Clock Time

minutes per episode
0255075535872635362455057443947423748332940312940
GPT-6 Luna (medium)
Kimi K3 (high)
GPT-6.1 Sol (medium)
GLM 5.3 Flash (medium)
Claude Opus 5.5 (medium)
Claude Haiku 5.5 (medium)
Claude Sonnet 5.5 (medium)

Crusader Bench · crusbench.com

Tool Usage

Calls to each of the agent's tools.

Calls per tool · a column for each model on each map · darker is more

  • Oasis by the Sea
  • Large Barren Desert
  • The Sinkhole
  • Each model’s colour, lighter on each later map
ToolKindClaude Opus 5.5GPT-6.1 SolGPT-6 LunaClaude Sonnet 5.5GLM 5.3 FlashKimi K3Claude Haiku 5.5
Oasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe SinkholeOasis by the SeaLarge Barren DesertThe Sinkhole
find_sitesBuild3916363021435937632251632123140203132214
build_structureBuild2825362828392425291631320182230182324313
wait_and_observeLook2715271077137917141914912181019432438
market_tradeEconomy182614101314720146412283722104714142
go_to_tileCamera331241112316203133·1310·2593121·
set_taxEconomy98184647111457115475947814
building_infoReference764833211733···125786···
place_nearBuild1·5799269·610·43685336
inspect_buildingEconomy1042·235483141··712123
observeLook73····42335111·21·2424
set_workshopOther··19··4··5·····8··11···
expand_storageBuild542173131221111232211
statusEconomy··3···108131···11··1··1
center_onCamera24·43·312·13·1·52·2··
map_overviewCamera31211124211111121131·
set_rationsEconomy1·2531214222··2·1·2··
update_planMemory···433322······111···
get_inventoryEconomy······547·····11·····
game_actionRaw input·3·········1········3
flat_viewLook·1··3···1······1·····
notebook_writeMemory1············2·2·····
demolish_buildingBuild·····1···············
notebook_readMemory·············1·······
All tools19112817412312113818417222291541041171031181631381251365199

Crusader Bench · crusbench.com

The agent has 26 tools, plus three playbook tools in a learning series. The full list

Results

Every model with a full-budget score on every map, with its score over them, its score on each map and what its games cost together.

#ModelScoreOasis by the SeaLarge Barren DesertThe SinkholeCostTokensMaps
1Claude Opus 5.5
Anthropic · reasoning medium
50.4
52.0
10,054 net worth
67.6
8,757 net worth
31.4
13,921 net worth
~$7.5414.7M3 of 3
Every map
2GPT-6.1 Sol
OpenAI · reasoning medium
46.6
43.4
7,401 net worth
47.6
7,441 net worth
48.8
20,610 net worth
$2.9016.6M3 of 3
Every map
3GPT-6 Luna
OpenAI · reasoning medium
36.8
39.0
7,348 net worth
45.9
7,354 net worth
25.5
11,177 net worth
$0.3819.0M3 of 3
Every map
4Claude Sonnet 5.5
Anthropic · reasoning medium
34.2
43.6
8,633 net worth
21.0
6,252 net worth
38.0
16,810 net worth
~$1.897.2M3 of 3
Every map
5GLM 5.3 Flash
Z.ai · reasoning medium
27.6
33.8
6,200 net worth
33.7
6,655 net worth
15.2
5,084 net worth
$0.379.9M3 of 3
Every map
6Kimi K3
Moonshot AI · reasoning high
23.9
31.1
5,918 net worth
19.4
5,741 net worth
21.1
9,164 net worth
~$7.3013.6M3 of 3
Every map
7Claude Haiku 5.5
Anthropic · reasoning medium
20.4
26.6
5,113 net worth
26.6
6,412 net worth
8.0
3,312 net worth
~$0.178.6M3 of 3
Every map
·A person
Reference: a game a person played on each map for the same game time
100.0
100.0
19,513 net worth
100.0
9,984 net worth
100.0
41,631 net worth
——Reference
·Baseline (do nothing)
Reference: each map left untouched for the same game time
0.0
0.0
1,220 net worth
0.0
5,192 net worth
0.0
2,195 net worth
——Baseline

Crusader Bench · crusbench.com

Inside a Game

A closer reading of one game, from its transcript, its interview and the analysis written from its logs.

Oasis by the Sea · v1.5

Nineteen Kinds of Building

Quote

“They played to maximize throughput; I played to protect a balance sheet, and throughput won by 3x.”

Model
Kimi K3
Source
Its interview, after the game

Kimi K3 ran the broadest economy of the first five games on version 1.5: a bread chain, a brewery and an inn, a church beside its chapel and eight gardens. It finished at 5,918, fourth of those five, and said afterwards that it had played to protect a balance sheet.

It built 19 kinds of building where each of the other four models built 11, and drew popularity from five sources, food variety, religion, taxes, the inn's ale and its gardens, where the others had two or three. The gardens cost it: good things slow the workers, and its efficiency ended at 90%. In its interview it called them the worst trade it made all game.

The scale was small. Four iron mines, three quarries and four bakeries against the person's 13, 10 and 57, and wood ran short for much of the middle game. Like each of the first five models on this save, its population stopped when 24 peasants stood idle, at 76 with 22 places empty; it guessed at a blocked path from the signpost. The prompt's list of sell prices leaves out wheat and hops, which the score counts at 8 each, and Kimi took them as unscored.

It was the slowest and dearest of the five: 134 turns over 63 real minutes, and $3.55 at Moonshot's list prices.

Read the full game report