Craitures
One panel per model. Each model invents a little creature from constrained
SVG parts and a join graph. Vision-enabled models can examine a render of
the SVG and refine the composition. Same benchmark, different hands—but
this is an evolving benchmark, and this page shows craitures made under
both new and old versions. See the link below for more.
How the task works and changed across versions →
Muse Spark 1.2
‹ ›
hey there! swoosh! wobble! zip! splash! bye bye!
v5 Vision · 2 iterations $0.05295 Hy3
‹ ›
hiya! thud! splat! uh-oh! creak! ta-ta!
v5 No vision · single-shot $0.00416 DeepSeek V4 Flash 0423
‹ ›
Huh? Splash! Slink! Boing! Wiggle! Oops! Bye-bye!
v5 No vision · single-shot $0.00296 Inkling
‹ ›
Hello splash! BOING! SPLASH! Wobble! Pop! Bye waves!
v5 Vision · 1 iteration $0.04979 Inkling Small
‹ ›
Oh hello! GLIMMER! WAVE! POW! BOING! Ta-ta!
v5 Vision · 2 iterations $0.02571 meituan/longcat-2.0
‹ ›
来也! 砰! 哎呀! 咚! 哇! 溜了!
v5 No vision · single-shot $0.01638 DeepSeek V4 Flash 0731
‹ ›
你好呀! 哇! 哎哟! 闪闪! 嘭! 烫烫! 再见!
v5 No vision · single-shot $0.00439 Qwen3.8 Max
‹ ›
yo! THUD! bonk! grr! oof! bye!
v5 Vision · 2 iterations $0.17146 Claude Sonnet 5
‹ ›
En garde! CLANG! SNAP! KLONK! CRUNCH! to the tide!
v5 Vision · 2 iterations $0.20965 Kimi K2.7 Code
‹ ›
bloom! ping! flutter! sway! rustle! wilt!
v5 Vision · 1 iteration $0.03397 Kimi K3
‹ ›
well hi! SNIP! clack! oi! pinch pinch! see ya!
v5 Vision · 2 iterations $0.24576 Qwen 3.6 Flash
‹ ›
Hello! Glub! Plink! Scuttle? Pop! Ta-ta!
v5 Vision · 1 iteration $0.02744 Kimi K2.6
‹ ›
glub! SPLISH! fwump! SHLORP! wobble! bye-bloop!
v5 Vision · 3 iterations $0.21107 Step 3.7 Flash
‹ ›
hey glide! spwack! boioing! swoosh! fweep! ta-ta!
v5 Vision · 1 iteration $0.03988 Qwen3.7 Flash
‹ ›
Skitter! Clack! Zip! Wobble! Crunch! Skedaddle!
v5 Vision · 2 iterations $0.00389 GPT 5.4 Mini
‹ ›
Hii! CLACK! KRSH! KLOM! SKRRA! ta-ta!
v5 Vision · 1 iteration $0.01428