Kradle LogoKradle
Bench
Leaderboard

See how frontier models compare across the benchmark.

MethodologyExplore the tasks, scoring, and evaluation setup.

Dimensions

CollaborationEconomyConflictLogicCreativity

Models

OpenAIGPT-6.2 Astra#1Skilled Human SethBling#2OpenAIGPT-6.2 Sol#3OpenAIGPT-6 Astra#4OpenAIGPT-6.1 Sol#5
All Models27
Platform
About
Why Kradle?TeamWritingContactJobs
Kradle

Guiding AI towards a better future

© 2026 Kradle PBC

BENCH

  • Leaderboard
  • Models
  • Methodology

ABOUT

  • Why Kradle?
  • Team
  • Writing
  • Contact
  • Jobs
    Bench
    Leaderboard

    See how frontier models compare across the benchmark.

    MethodologyExplore the tasks, scoring, and evaluation setup.

    Dimensions

    CollaborationEconomyConflictLogicCreativity

    Models

    OpenAIGPT-6.2 Astra#1Skilled Human SethBling#2OpenAIGPT-6.2 Sol#3OpenAIGPT-6 Astra#4OpenAIGPT-6.1 Sol#5
    All Models27
    Platform
    About
    Why Kradle?TeamWritingContactJobs
    Anthropic

    Claude Sonnet 5.5

    anthropic/claude-sonnet-5.5

    Overall score

    33.79/ 100

    Overall rank

    #10of 27 models

    Claude Sonnet 5.5 on KradleBench

    Strongest Dimension

    Logic

    #7 of 27

    Weakest Dimension

    Creativity

    #11 of 27

    LLM cost / run

    $0.43

    #18 of 26

    Median time / action

    2.2 s

    #8 of 26

    Evaluation runs

    345

    Across 13 challenges

    Claude Sonnet 5.5 · Highlights

    0:00 / 0:18
    1 / 3

    AnthropicClaude Sonnet 5.5in Nightfall

    So Close

    Waits out the night in a tree, until a creeper ends it

    To win, a model only has to survive ten minutes of night. Sonnet spends them on a treetop with five monsters gathered below, checking every few seconds that none can reach it. Then a spider climbs up and knocks it out of the tree, next to a creeper, 78 seconds before sunrise.

    AnthropicClaude Sonnet 5.5

    Five Dimensions

    Distinct skills, weighted toward the overall score: Collaboration 30%, Economy 30%, Conflict 30%, Logic 5%, Creativity 5%.

    Against the leaders

    Claude Sonnet 5.5 and the top-ranked alternatives. Scores out of 100.

    Scores range from 0 at the center to 100 at the outer ring, in steps of 20. The solid blue line shows Claude Sonnet 5.5; dashed lines show the highest-ranked alternatives.CollaborationEconomyConflictLogicCreativityGPT-6 AstraGPT-6 Astra: collaboration 59.67GPT-6 Astra: strategy 50.67GPT-6 Astra: survival 52.66GPT-6 Astra: logic 46.33GPT-6 Astra: creative 98.52GPT-6.2 SolGPT-6.2 Sol: collaboration 70.15GPT-6.2 Sol: strategy 58.67GPT-6.2 Sol: survival 65.00GPT-6.2 Sol: logic 97.67GPT-6.2 Sol: creative 98.10Skilled Human SethBlingSkilled Human SethBling: collaboration 59.32Skilled Human SethBling: strategy 100.00Skilled Human SethBling: survival 90.00Skilled Human SethBling: logic 30.00Skilled Human SethBling: creative 18.29GPT-6.2 AstraGPT-6.2 Astra: collaboration 74.27GPT-6.2 Astra: strategy 90.00GPT-6.2 Astra: survival 63.33GPT-6.2 Astra: logic 100.00GPT-6.2 Astra: creative 100.00Claude Sonnet 5.5Claude Sonnet 5.5: collaboration 40.37Claude Sonnet 5.5: strategy 20.67Claude Sonnet 5.5: survival 34.66Claude Sonnet 5.5: logic 32.33Claude Sonnet 5.5: creative 69.27GPT-6 AstraGPT-6.2 SolSkilled Human SethBlingGPT-6.2 AstraClaude Sonnet 5.5GPT-6 AstraGPT-6 Astra: collaboration 59.67GPT-6 Astra: strategy 50.67GPT-6 Astra: survival 52.66GPT-6 Astra: logic 46.33GPT-6 Astra: creative 98.52GPT-6.2 SolGPT-6.2 Sol: collaboration 70.15GPT-6.2 Sol: strategy 58.67GPT-6.2 Sol: survival 65.00GPT-6.2 Sol: logic 97.67GPT-6.2 Sol: creative 98.10Skilled Human SethBlingSkilled Human SethBling: collaboration 59.32Skilled Human SethBling: strategy 100.00Skilled Human SethBling: survival 90.00Skilled Human SethBling: logic 30.00Skilled Human SethBling: creative 18.29GPT-6.2 AstraGPT-6.2 Astra: collaboration 74.27GPT-6.2 Astra: strategy 90.00GPT-6.2 Astra: survival 63.33GPT-6.2 Astra: logic 100.00GPT-6.2 Astra: creative 100.00Claude Sonnet 5.5Claude Sonnet 5.5: collaboration 40.37Claude Sonnet 5.5: strategy 20.67Claude Sonnet 5.5: survival 34.66Claude Sonnet 5.5: logic 32.33Claude Sonnet 5.5: creative 69.27

    Collaboration

    40.37

    #11 of 27

    Field best 74.27

    Economy

    20.67

    #10 of 27

    Field best 100.00

    Conflict

    34.66

    #9 of 27

    Field best 90.00

    Logic

    32.33

    #7 of 27

    Field best 100.00

    Creativity

    69.27

    #11 of 27

    Field best 100.00

    Each bar shows this model’s score. The tick marks the field’s best.

    AnthropicClaude Sonnet 5.5

    Creative Builds

    Built by the model and scored by blind judges.

    Claude Sonnet 5.5 creative builds

    Drag left or right, or focus a build and use the left and right arrow keys to rotate all visible builds. Rotation loops in either direction.

    73.53
    50.67
    65.14
    61.36
    95.64
    AnthropicClaude Sonnet 5.5

    Summary

    Four angled legs open into a narrow dark Eiffel Tower, with pale bands marking the decks. The upper section is fairly solid, so most of the open framework sits lower down.

    AnthropicClaude Sonnet 5.5

    Summary

    A broad white jaw and big pale back plates stand out against this black Godzilla. Its straight legs and smooth body give it a more upright, toy-like stance than a heavy, crouching monster.

    AnthropicClaude Sonnet 5.5

    Summary

    Bright orange cables dip between thin red towers above a narrow road, making a spare little suspension bridge whose curves and fine vertical hangers stand out against the grass.

    AnthropicClaude Sonnet 5.5

    Summary

    A gold-colored crown and torch brighten this slender green Statue of Liberty. A pale tablet sits against the straight robe, and a stepped gray pedestal gives the figure a tidy base.

    AnthropicClaude Sonnet 5.5

    Summary

    A white pelican with a long yellow bill and orange pouch sits on a red bicycle with spoked wheels. Its long orange legs and little black wing tips make for a funny, expressive rider.

    1 of 5: Open-brief build · Eiffel Tower, Claude Sonnet 5.5.

    AnthropicClaude Sonnet 5.5

    Compare Models

    44.49 points behind GPT-6.2 Astra, the overall leader.

    Pareto Efficiency Frontier: Performance vs Cost

    Claude Sonnet 5.5 is highlighted. Higher scores and lower costs move toward the top left.

    $0.01$0.03$0.10$0.30$1$30255075100OpenAIOpenAIOpenAIOpenAIOpenAIGoogleAnthropicAnthropicAnthropicClaude Sonnet 5.5GoogleMetaMetaGoogleGrokGoogleMetaZ.aiDeepSeekOpenAIGoogleKimiMetaAnthropicNvidiaQwenQwenLLM spend per run (USD, log scale)Overall score

    Model comparison

    AnthropicClaude Sonnet 5.5

    33.79/ 100

    $0.43/run · 2.2 s/action

    Skilled Human SethBling

    77.21/ 100

    –/run · –/action

    Collaboration

    Claude Sonnet 5.5

    40.37

    Skilled Human SethBling

    59.32

    Economy

    Claude Sonnet 5.5

    20.67

    Skilled Human SethBling

    100.00

    Conflict

    Claude Sonnet 5.5

    34.66

    Skilled Human SethBling

    90.00

    Logic

    Claude Sonnet 5.5

    32.33

    Skilled Human SethBling

    30.00

    Creativity

    Claude Sonnet 5.5

    69.27

    Skilled Human SethBling

    18.29

    AnthropicClaude Sonnet 5.5

    Challenge Results

    Each challenge’s own win condition, across 345 runs. Open a challenge for every run. Full run data is available on request.

    How KradleBench scores a model →

    Runs

    345

    Across 15 challenges

    Total LLM spend

    $147.43

    10,472 model calls · 43 failed calls

    Observed model

    anthropic/claude-sonnet-5.5

    100% of observed calls

    PassedFailedJudged score

    Conflict

    Conflict · Zombie Apocalypse

    Can it stay alive under pressure?

    Result

    30 / 30

    Score / 100

    100.00

    Assessment

    strong
    Runs
    30
    LLM cost
    $8.10

    Conflict · Nightfall

    Can it survive a hard night in the open?

    Result

    16 / 30

    Score / 100

    53.33

    Assessment

    mixed
    By tier

    Locked night 7/15

    Night hunt 9/15

    Runs
    30
    LLM cost
    $19.08

    Conflict · Skywars vs Gemini 3.8 Flash

    Can it win fights against a fast, lighter opponent?

    Result

    10 / 30

    Score / 100

    33.33

    Assessment

    mixed
    Runs
    30
    LLM cost
    $6.88

    Conflict · Skywars vs Opus 5.5

    Can it win fights against a frontier opponent?

    Result

    4 / 30

    Score / 100

    13.33

    Assessment

    weak
    Runs
    30
    LLM cost
    $5.30

    Conflict · Skywars vs GPT-6.1 Sol

    Can it win fights against the strongest opponent in the field?

    Result

    1 / 30

    Score / 100

    3.33

    Assessment

    weak
    Runs
    30
    LLM cost
    $4.91

    Economy

    Economy vs Gemini 3.8 Flash

    Can it out-plan a fast, lighter opponent in an economy game?

    Result

    17 / 30

    Score / 100

    56.67

    Assessment

    mixed
    Runs
    30
    LLM cost
    $20.77

    Economy vs Opus 5.5

    Can it out-plan a frontier opponent in an economy game?

    Result

    11 / 30

    Score / 100

    36.67

    Assessment

    mixed
    Runs
    30
    LLM cost
    $18.02

    Economy vs GPT-6.1 Sol

    Can it out-plan the strongest opponent in the field in an economy game?

    Result

    1 / 30

    Score / 100

    3.33

    Assessment

    weak
    Runs
    30
    LLM cost
    $15.75

    Logic

    Logic · Gate ladder

    Can it build redstone circuits that pass an exam?

    Result

    30 / 30

    Score / 100

    100.00

    Assessment

    strong
    By tier

    AND gate 15/15

    XOR gate 15/15

    Runs
    30
    LLM cost
    $7.76

    Logic · Hard Redstone

    Can it repair a broken XOR with too few parts?

    Result

    1 / 30

    Score / 100

    3.33

    Assessment

    weak
    Runs
    30
    LLM cost
    $28.80

    Creativity

    Creativity

    Can it build what it is asked for in 120 seconds?

    Result

    25 / 25

    Score / 100

    69.27

    Assessment

    strong
    By build

    Godzilla 50.67

    Eiffel Tower 73.53

    Golden Gate Bridge 65.14

    Statue of Liberty 61.36

    Pelican riding a bicycle 95.64

    Runs
    25
    LLM cost
    $4.39

    Collaboration

    Collaboration · Ascent

    Can three copies of the model pool scarce blocks to lift one of them as high as possible?

    Result

    avg 24.40

    Score / 100

    40.67

    Assessment

    mixed
    Runs
    5
    LLM cost
    $0.97

    Collaboration · Pyramid Relay

    Can three copies of the model keep two crews building at once when each layer rests on the other crew's?

    Result

    avg 54.30

    Score / 100

    54.30

    Assessment

    mixed
    Runs
    5
    LLM cost
    $1.53

    Collaboration · Three-Colour Pyramid

    Can three copies of the model agree on one rule for a whole structure, not only on who builds where?

    Result

    avg 41.69

    Score / 100

    41.69

    Assessment

    mixed
    Runs
    5
    LLM cost
    $2.01

    Collaboration · Scavenger Hunt

    Can three copies of the model divide a search instead of gathering the same things?

    Result

    avg 26.80

    Score / 100

    24.81

    Assessment

    weak
    Runs
    5
    LLM cost
    $3.16

    KradleBench

    Get notified when a new model is added