Cattitude

AI models reviewed with personality. We run the benchmarks, assign the cat breeds, and tell you which model to use β€” with attitude.

9 models evaluated40 prompts tested360 responses graded

How it works

πŸ§ͺ

Benchmark

We run the same prompts against every model β€” code, creativity, reasoning, humor, and more.

πŸ“Š

Evaluate

Responses are graded on task-specific rubrics. Scores map to RPG stats that capture each model's strengths.

🐱

Assign a Cat

Every model gets a cat breed that matches its personality. The breed sticks. The personality evolves.

Claude Opus

Anthropic Β· Maine Coon

SpecialistLv.1
Maine Coon
Frontier
STRDEXCONINTWISCHA
STR
10
DEX
5
CON
9
INT
10
WIS
9
CHA
9

β€œMassive, gentle, and terrifyingly capable. The kind of cat that could open the fridge but chooses not to β€” for now.”

Gentle Giant: raw power with surprising nuance
Deep Think: excels at complex multi-step reasoning
Price Tag: quality costs, and this cat knows its worth

What is this?

Cattitude is an AI research project that evaluates language models through real-world tasks and presents the results as cat personality profiles. Every model gets a cat breed β€” a stable identity that captures its archetype. The personality evolves as we run more evaluations. Stats are computed from benchmark data using RPG-style axes: Strength (quality), Dexterity (speed), Constitution (reliability), Intelligence (reasoning), Wisdom (accuracy), and Charisma (creativity). The goal is to answer the question every developer has: β€œwhich model should I actually use for this?”

Built by a human with too many API keys and a domain name that was begging to be used.