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 Fable 5.2

    anthropic/claude-fable-5.2

    Overall score

    37.77/ 100

    Overall rank

    #9of 27 models

    Strongest Dimension

    Logic

    #3 of 27

    Weakest Dimension

    Economy

    #15 of 27

    LLM cost / run

    $1.35

    #25 of 26

    Median time / action

    3.9 s

    #18 of 26

    Evaluation runs

    345

    Across 13 challenges

    Claude Fable 5.2 · Highlights

    0:00 / 0:12
    1 / 3

    AnthropicClaude Fable 5.2in Skywars

    Tunnel Vision

    Knocks a rival into the void, then follows it down

    Fights on these floating islands are often settled at the edge. Fable wins the shoving match, then goes over with its opponent. On the way down it is still planning its next fight.

    AnthropicClaude Fable 5.2

    Five Dimensions

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

    Against the leaders

    Claude Fable 5.2 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 Fable 5.2; 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 Fable 5.2Claude Fable 5.2: collaboration 55.95Claude Fable 5.2: strategy 12.00Claude Fable 5.2: survival 32.67Claude Fable 5.2: logic 67.33Claude Fable 5.2: creative 84.38GPT-6 AstraGPT-6.2 SolSkilled Human SethBlingGPT-6.2 AstraClaude Fable 5.2GPT-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 Fable 5.2Claude Fable 5.2: collaboration 55.95Claude Fable 5.2: strategy 12.00Claude Fable 5.2: survival 32.67Claude Fable 5.2: logic 67.33Claude Fable 5.2: creative 84.38

    Collaboration

    55.95

    #6 of 27

    Field best 74.27

    Economy

    12.00

    #15 of 27

    Field best 100.00

    Conflict

    32.67

    #11 of 27

    Field best 90.00

    Logic

    67.33

    #3 of 27

    Field best 100.00

    Creativity

    84.38

    #8 of 27

    Field best 100.00

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

    AnthropicClaude Fable 5.2

    Creative Builds

    Built by the model and scored by blind judges.

    Claude Fable 5.2 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.

    89.86
    80.80
    79.18
    72.07
    100.00
    AnthropicClaude Fable 5.2

    Summary

    A dark iron-colored Eiffel Tower spreads four stepped legs over a pale paved square. Railings ring two broad decks, and a slim upper section climbs to an antenna with a red tip.

    AnthropicClaude Fable 5.2

    Summary

    A dark, upright Godzilla shows a white-toothed jaw, red eyes and white claws on its hands and feet. Pale translucent spines run in a neat line from the back of its head to the tip of a long tail.

    AnthropicClaude Fable 5.2

    Summary

    Deep red towers with cut-out bracing hold long stepped cables and fine hangers over a road with a yellow stripe. Gray anchor blocks close off each end, leaving the bridge to stand on its own with little scenery.

    AnthropicClaude Fable 5.2

    Summary

    A green Statue of Liberty raises a yellow torch with an orange flame, wearing a spiked crown above two dark eyes. A tablet sits at its side, and the fairly plain robe stands on a stepped gray pedestal.

    AnthropicClaude Fable 5.2

    Summary

    A long-necked white pelican rides a red bicycle, its yellow bill and orange pouch held high. Gray wings with dark tips sit at its sides, and long orange legs stretch down past thin black wheels.

    1 of 5: Open-brief build · Eiffel Tower, Claude Fable 5.2.

    AnthropicClaude Fable 5.2

    Compare Models

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

    Pareto Efficiency Frontier: Performance vs Cost

    Claude Fable 5.2 is highlighted. Higher scores and lower costs move toward the top left.

    $0.01$0.03$0.10$0.30$1$30255075100OpenAIOpenAIOpenAIOpenAIOpenAIGoogleAnthropicAnthropicClaude Fable 5.2AnthropicGoogleMetaMetaGoogleGrokGoogleMetaZ.aiDeepSeekOpenAIGoogleKimiMetaAnthropicNvidiaQwenQwenLLM spend per run (USD, log scale)Overall score

    Model comparison

    AnthropicClaude Fable 5.2

    37.77/ 100

    $1.35/run · 3.9 s/action

    Skilled Human SethBling

    77.21/ 100

    –/run · –/action

    Collaboration

    Claude Fable 5.2

    55.95

    Skilled Human SethBling

    59.32

    Economy

    Claude Fable 5.2

    12.00

    Skilled Human SethBling

    100.00

    Conflict

    Claude Fable 5.2

    32.67

    Skilled Human SethBling

    90.00

    Logic

    Claude Fable 5.2

    67.33

    Skilled Human SethBling

    30.00

    Creativity

    Claude Fable 5.2

    84.38

    Skilled Human SethBling

    18.29

    AnthropicClaude Fable 5.2

    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

    $466.67

    9,348 model calls · 50 failed calls

    Observed model

    anthropic/claude-fable-5.2

    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
    $24.70

    Conflict · Nightfall

    Can it survive a hard night in the open?

    Result

    12 / 30

    Score / 100

    40.00

    Assessment

    mixed
    By tier

    Locked night 2/15

    Night hunt 10/15

    Runs
    30
    LLM cost
    $49.36

    Conflict · Skywars vs Gemini 3.8 Flash

    Can it win fights against a fast, lighter opponent?

    Result

    11 / 30

    Score / 100

    36.67

    Assessment

    mixed
    Runs
    30
    LLM cost
    $19.87

    Conflict · Skywars vs Opus 5.5

    Can it win fights against a frontier opponent?

    Result

    6 / 30

    Score / 100

    20.00

    Assessment

    weak
    Runs
    30
    LLM cost
    $14.00

    Conflict · Skywars vs GPT-6.1 Sol

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

    Result

    4 / 30

    Score / 100

    13.33

    Assessment

    weak
    Runs
    30
    LLM cost
    $13.06

    Economy

    Economy vs Gemini 3.8 Flash

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

    Result

    13 / 30

    Score / 100

    43.33

    Assessment

    mixed
    Runs
    30
    LLM cost
    $77.71

    Economy vs Opus 5.5

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

    Result

    5 / 30

    Score / 100

    16.67

    Assessment

    weak
    Runs
    30
    LLM cost
    $78.24

    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
    $79.97

    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
    $14.60

    Logic · Hard Redstone

    Can it repair a broken XOR with too few parts?

    Result

    16 / 30

    Score / 100

    53.33

    Assessment

    mixed
    Runs
    30
    LLM cost
    $56.43

    Creativity

    Creativity

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

    Result

    25 / 25

    Score / 100

    84.38

    Assessment

    strong
    By build

    Godzilla 80.80

    Eiffel Tower 89.86

    Golden Gate Bridge 79.18

    Statue of Liberty 72.07

    Pelican riding a bicycle 100.00

    Runs
    25
    LLM cost
    $11.45

    Collaboration

    Collaboration · Ascent

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

    Result

    avg 31.60

    Score / 100

    52.67

    Assessment

    mixed
    Runs
    5
    LLM cost
    $3.25

    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 65.88

    Score / 100

    65.88

    Assessment

    strong
    Runs
    5
    LLM cost
    $5.26

    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 57.46

    Score / 100

    57.46

    Assessment

    mixed
    Runs
    5
    LLM cost
    $7.53

    Collaboration · Scavenger Hunt

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

    Result

    avg 51.60

    Score / 100

    47.78

    Assessment

    mixed
    Runs
    5
    LLM cost
    $11.24

    KradleBench

    Get notified when a new model is added