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 →
Mistral Large 4
‹ ›
drift ho! blub blub! PLOP! shimmer! bob bob! bye-bye!
v7 Vision · 1 iteration $0.00969 Solar Mini 4
‹ ›
ooh! yarr! SPLASH! CLANK! pop! SPLAY! SHIP! via the reef ahoy!
v7 No vision · single-shot $0.00081 DeepSeek V4.1 Flash
‹ ›
hello sun! BOING! creak-rustle! petal-pop! root-thump! ta-ta bees!
v7 Vision · 3 iterations $0.1211 Apodex 1.1 Mini (free)
‹ ›
Hello! Scritch! Clack! Pop! Grrr! Ta-ta!
v7 No vision · single-shot $0.00 Ling 3.1 Flash
‹ ›
hiya puddle! SPLAT! blurp! squish! whoosh! ta-ta mud!
v7 No vision · single-shot $0.00 Pareto 26.10 Preview
‹ ›
Kiln hello! CLONK! Tink tank! SKONK! Hot hop! Stride away!
v7 Vision · 1 iteration $0.06195 Grok 4.7
‹ ›
glow hi! TING! ay! BONK! spark! dim bye!
v7 Vision · 2 iterations $0.14918 GPT-6.1 Sol
‹ ›
Ahoy! Rattle! Thunk! Whuff! Skrrt! Fair winds!
v7 Vision · 1 iteration $0.03587 GPT-6.1 Sol Pro
‹ ›
Ahoy! Flap! Clack! Whoosh! Bonk! Sail on!
v7 Vision · 1 iteration $0.10844 Claude Sonnet 5.5
‹ ›
hullo! CLONK! THUD! BONK! CLANG! plod on!
v7 Vision · 1 iteration $0.06272 GPT-6 Luna
‹ ›
Wheel hello! WHUM! TINK! FWOOP! CLANG! Drift on!
v7 Vision · 2 iterations $0.00372 Command A+
‹ ›
Hey there! Zap! Wham! Boom! Whoosh! See ya!
v7 Vision · 2 iterations $0.11425 Space Bunny Alpha
‹ ›
Hi husk! CLAK! KRUMP! SKIT! ZING! Later chitin!
v7 Vision · 2 iterations $0.00 Ember-1
‹ ›
ahoy-oi! FWOOSH! bonk! wheee! swoop! glide on!
v7 Vision · 2 iterations $0.14574 MiMo-V2.6-Pro
‹ ›
hello sprout! WHIRR! CLANK! skiff-skid! boing! seed you later!
v7 Vision · 3 iterations $0.03638 Granite 4.2 8B
‹ ›
Hello seed! POP! BOING! ZAP! WHIRL! Bye seed!
v7 No vision · single-shot $0.00289