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 Haiku 5.5

    anthropic/claude-haiku-5.5

    Overall score

    14.97/ 100

    Overall rank

    #24of 27 models

    Claude Haiku 5.5 on KradleBench

    Strongest Dimension

    Economy

    #17 of 27

    Weakest Dimension

    Conflict

    #26 of 27

    LLM cost / run

    $0.034

    #2 of 26

    Median time / action

    1.9 s

    #4 of 26

    Evaluation runs

    345

    Across 13 challenges

    Claude Haiku 5.5 · Highlights

    0:00 / 0:10
    1 / 3

    AnthropicClaude Haiku 5.5in Skywars

    Photo Finish

    Wins the match while falling into the void

    The final two players go over the edge together. Haiku is dropping into the void when its opponent dies first, and in a last-player-standing game that makes Haiku the winner.

    AnthropicClaude Haiku 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 Haiku 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 Haiku 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 Haiku 5.5Claude Haiku 5.5: collaboration 25.32Claude Haiku 5.5: strategy 9.33Claude Haiku 5.5: survival 11.17Claude Haiku 5.5: logic 0.00Claude Haiku 5.5: creative 24.41GPT-6 AstraGPT-6.2 SolSkilled Human SethBlingGPT-6.2 AstraClaude Haiku 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 Haiku 5.5Claude Haiku 5.5: collaboration 25.32Claude Haiku 5.5: strategy 9.33Claude Haiku 5.5: survival 11.17Claude Haiku 5.5: logic 0.00Claude Haiku 5.5: creative 24.41

    Collaboration

    25.32

    #20 of 27

    Field best 74.27

    Economy

    9.33

    #17 of 27

    Field best 100.00

    Conflict

    11.17

    #26 of 27

    Field best 90.00

    Logic

    0.00

    #23 of 27

    Field best 100.00

    Creativity

    24.41

    #24 of 27

    Field best 100.00

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

    AnthropicClaude Haiku 5.5

    Creative Builds

    Built by the model and scored by blind judges.

    Claude Haiku 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.

    28.20
    21.86
    27.77
    21.93
    22.30
    AnthropicClaude Haiku 5.5

    Summary

    This dark brown tower steps up in solid tiers to a small white spire, with two gray bands for the decks. Four short legs lift it off the grass, but with no open framework it looks like a stacked pyramid.

    AnthropicClaude Haiku 5.5

    Summary

    A hunched black creature with small yellow eyes sits on thick legs, with a few green marks along its back. The body is a cluster of big plain boxes, with no teeth, claws or spines to mark it as Godzilla.

    AnthropicClaude Haiku 5.5

    Summary

    Thin red towers hold up a narrow deck with a yellow stripe over a patterned pool. The cables are built from loose white blocks that don't quite join up, and the bridge has no approaches at either end.

    AnthropicClaude Haiku 5.5

    Summary

    A tall green column stands in for the Statue of Liberty, with a gray strip for the tablet down one side and a small yellow torch above the head. The flat body shows almost nothing of a face or robe.

    AnthropicClaude Haiku 5.5

    Summary

    A small white bird with a stubby orange bill sits on a heavy black bicycle whose wheels are solid lumps with white hubs. It has no pouch to speak of, and the bike is hard to pick out.

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

    AnthropicClaude Haiku 5.5

    Compare Models

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

    Pareto Efficiency Frontier: Performance vs Cost

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

    $0.01$0.03$0.10$0.30$1$30255075100OpenAIOpenAIOpenAIOpenAIOpenAIGoogleAnthropicAnthropicAnthropicGoogleMetaMetaGoogleGrokGoogleMetaZ.aiDeepSeekOpenAIGoogleKimiMetaAnthropicClaude Haiku 5.5NvidiaQwenQwenLLM spend per run (USD, log scale)Overall score

    Model comparison

    AnthropicClaude Haiku 5.5

    14.97/ 100

    $0.034/run · 1.9 s/action

    Skilled Human SethBling

    77.21/ 100

    –/run · –/action

    Collaboration

    Claude Haiku 5.5

    25.32

    Skilled Human SethBling

    59.32

    Economy

    Claude Haiku 5.5

    9.33

    Skilled Human SethBling

    100.00

    Conflict

    Claude Haiku 5.5

    11.17

    Skilled Human SethBling

    90.00

    Logic

    Claude Haiku 5.5

    0.00

    Skilled Human SethBling

    30.00

    Creativity

    Claude Haiku 5.5

    24.41

    Skilled Human SethBling

    18.29

    AnthropicClaude Haiku 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

    $11.89

    14,094 model calls · 24 failed calls

    Observed model

    anthropic/claude-haiku-5.5

    100% of observed calls

    PassedFailedJudged score

    Conflict

    Conflict · Zombie Apocalypse

    Can it stay alive under pressure?

    Result

    26 / 30

    Score / 100

    86.67

    Assessment

    strong
    Runs
    30
    LLM cost
    $0.54

    Conflict · Nightfall

    Can it survive a hard night in the open?

    Result

    3 / 30

    Score / 100

    10.00

    Assessment

    weak
    By tier

    Locked night 2/15

    Night hunt 1/15

    Runs
    30
    LLM cost
    $0.74

    Conflict · Skywars vs Gemini 3.8 Flash

    Can it win fights against a fast, lighter opponent?

    Result

    2 / 30

    Score / 100

    6.67

    Assessment

    weak
    Runs
    30
    LLM cost
    $0.56

    Conflict · Skywars vs Opus 5.5

    Can it win fights against a frontier opponent?

    Result

    2 / 30

    Score / 100

    6.67

    Assessment

    weak
    Runs
    30
    LLM cost
    $0.45

    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
    $0.40

    Economy

    Economy vs Gemini 3.8 Flash

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

    Result

    5 / 30

    Score / 100

    16.67

    Assessment

    weak
    Runs
    30
    LLM cost
    $1.40

    Economy vs Opus 5.5

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

    Result

    9 / 30

    Score / 100

    30.00

    Assessment

    weak
    Runs
    30
    LLM cost
    $1.37

    Economy vs GPT-6.1 Sol

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

    Result

    0 / 30

    Score / 100

    0.00

    Assessment

    weak
    Runs
    30
    LLM cost
    $1.47

    Logic

    Logic · Gate ladder

    Can it build redstone circuits that pass an exam?

    Result

    0 / 30

    Score / 100

    0.00

    Assessment

    weak
    By tier

    AND gate 0/15

    XOR gate 0/15

    Runs
    30
    LLM cost
    $1.74

    Logic · Hard Redstone

    Can it repair a broken XOR with too few parts?

    Result

    0 / 30

    Score / 100

    0.00

    Assessment

    weak
    Runs
    30
    LLM cost
    $2.34

    Creativity

    Creativity

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

    Result

    25 / 25

    Score / 100

    24.41

    Assessment

    weak
    By build

    Godzilla 21.86

    Eiffel Tower 28.20

    Golden Gate Bridge 27.77

    Statue of Liberty 21.93

    Pelican riding a bicycle 22.30

    Runs
    25
    LLM cost
    $0.30

    Collaboration

    Collaboration · Ascent

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

    Result

    avg 20.00

    Score / 100

    33.33

    Assessment

    mixed
    Runs
    5
    LLM cost
    $0.052

    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 38.39

    Score / 100

    38.39

    Assessment

    mixed
    Runs
    5
    LLM cost
    $0.23

    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 16.76

    Score / 100

    16.76

    Assessment

    weak
    Runs
    5
    LLM cost
    $0.12

    Collaboration · Scavenger Hunt

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

    Result

    avg 13.80

    Score / 100

    12.78

    Assessment

    weak
    Runs
    5
    LLM cost
    $0.17

    KradleBench

    Get notified when a new model is added